How to Write an NSFC (National Natural Science Foundation of China) Proposal: Section-by-Section Guide
2026/11/16

How to Write an NSFC (National Natural Science Foundation of China) Proposal: Section-by-Section Guide

A section-by-section guide to the 2026 NSFC General Program outline, locked headings, Young Scientists Fund (A/B/C) rules, and AI disclosure policy.

Short answer: an NSFC (National Natural Science Foundation of China, 国家自然科学基金) proposal is neither a broad literature review nor an expanded draft of a journal article, but a focused feasibility argument organized around one to three key scientific questions (关键科学问题). Under the 2026 General Program (面上项目) body outline, the main text should in principle stay within 30 pages (including figures, tables, and references), and none of the official Chinese headings or parenthetical instructions may be deleted or altered: Part I ((一)立项依据) establishes the scientific problem and its value, Part II ((二)研究内容) connects objectives, methods, innovations, and the annual schedule according to your project's internal logic, and Part III ((三)研究基础) uses preliminary findings, risk contingencies, and institutional facilities to show that your team can execute the work.

Three rigid compliance rules to verify before writing

For international faculty, postdoctoral fellows, and overseas-trained scholars returning to host institutions in China, preparing an NSFC General Program (面上项目) application differs structurally from NIH or European grant formats because the main body must be submitted inside a Chinese Word template with locked outline text. Before drafting, always check the current year's NSFC Program Guide (《国家自然科学基金项目指南》), annual application notice, and official Word template downloaded from the NSFC online system (科学基金网络信息系统). Based on the 2026 Annual Notice (国科金发计〔2026〕2号, see university-forwarded notice attachment) and the 2026 official Word template archive, three compliance rules govern preliminary administrative screening (形式审查):

  1. 30-page length ceiling: The official body template states "申请书正文原则上不超过 30 页,鼓励简洁表达" (The main body of the proposal should in principle not exceed 30 pages; concise expression is encouraged), inclusive of technical roadmaps, preliminary data figures and tables, and references.
  2. Locked headings and parenthetical instructions: The official template stipulates "请勿删除或改动下述提纲标题及括号中的文字" (Do not delete or modify the outline headings below or the text in parentheses). Whenever a subsection does not apply to your situation, keep the original Chinese heading and parenthetical text intact and enter 无 (None) beneath it.
  3. Research attribute and category-specific review: In the online form, applicants must select either 自由探索类基础研究 (curiosity-driven basic research) or 目标导向类基础研究 (goal-oriented basic research). Notice 国科金发计〔2026〕2号 specifies that for programs piloting category-specific review—including the General Program (面上项目), Young Scientists Fund Category C (青年科学基金项目(C类)), and Key Program (重点项目)—NSFC organizes peer review according to the selected research attribute. Your opening framing in the body text must match that choice.

The 2026 General Program body consists of four top-level parts: (一)立项依据 (Rationale and Justification), (二)研究内容 (Research Content), (三)研究基础 (Research Foundation), and (四)其他需要说明的情况 (Other Matters Requiring Explanation). Note that the Regional Science Fund (地区科学基金项目) in 2026 retains the older three-part outline, with five subsections under Part I and a suggested length under 8,000 Chinese characters; if future outlines are updated, follow that year's official NSFC template. NSFC does not mandate a uniform body font or size; according to the Xihua University annotated 2026 General Program sample, applicants should not insert headers, footers, or page numbers while editing, and should paste the finalized text into the editing-restricted official Word template downloaded from the online system before uploading.

Part / Form FieldOfficial 2026 General Program Instruction (Chinese + English Gloss)Core Writing TaskLength & Formatting Notes
Bilingual Abstracts (中英文摘要)Entered in the online system formCondense background, key scientific question, content, methods, and scientific valueInstitutional sample recommends Chinese abstract within 400 characters (including punctuation); no figures, tables, or equations
(一)立项依据 (Part I: Rationale and Justification)(为什么要开展此项研究,研究的科学技术价值如何) (Why conduct this research, and what is its scientific and technological value?)Establish the origin of the problem, knowledge gap, and necessity of the workIntroduce key scientific questions in the second half; place the main reference list immediately at the end of Part I
(二)研究内容 (Part II: Research Content)(提纲不做限制,请按照研究工作的自身逻辑撰写。应提炼出特色与创新点、年度研究计划) (Outline unrestricted; write according to the internal logic of your research, distilling distinctive features & innovations and the annual plan)Present objectives, content, key scientific questions, study design, innovations, annual plan, and expected outcomesKeep innovations to no more than 3 points (recommended); annual plan dates must match the project period exactly
(三)研究基础 (Part III: Research Foundation)Contains 4 locked subsections (feasibility & risks, working conditions, ongoing grants, prior NSFC grant completion)Substantiate feasibility via prior work, preliminary data, risk contingencies, and facilitiesSubsection 4 requires a work summary abstract of the preceding completed NSFC grant (capped at 500 Chinese characters)
(四)其他需要说明的情况 (Part IV: Other Matters)Contains 5 locked subsections (same-year NSFC submissions, institutional mismatches, applications under different titles, etc.)Compliance disclosures on submission limits and institutional affiliation consistencyRetain every heading and enter 无 (None) where inapplicable

Abstracts and Part I: (一)立项依据 (Rationale and Justification)

Bilingual abstracts (中文摘要 and 英文摘要)

Although entered in the online form rather than the Word attachment, the abstracts are the first window mail reviewers (函评专家) read. Following the annotated institutional sample, keep the Chinese abstract within 400 Chinese characters (including punctuation) and structure it as a tight five-sentence loop: (1) research object and bottleneck, (2) key scientific question or central hypothesis, (3) main research content, (4) core methodology, and (5) mechanism to be revealed and its scientific value. Ensure the English abstract corresponds accurately in meaning.

(一)立项依据 (Rationale and Justification): what to cover and common pitfalls

The 2026 official parenthetical instruction reads (为什么要开展此项研究,研究的科学技术价值如何) (Why conduct this research, and what is its scientific and technological value?). Even with this streamlined prompt, avoid turning Part I into a textbook-style survey. A clear six-step progression works well for specialist reviewers:

  1. Open from a concrete necessity: Start with a specific scientific anomaly or engineering bottleneck, framing the problem and scope within two or three paragraphs.
  2. Critique core literature: Trace international and domestic advances along the evolution of the problem, identifying where existing theories or methods break down.
  3. Converge on the knowledge gap: Pinpoint the exact mechanistic or methodological link your project will resolve.
  4. Articulate scientific and technological value: Explain how resolving that link advances disciplinary theory or key engineering capability.
  5. Lead into the key scientific questions: In the latter half of Part I, state your central hypothesis and key scientific questions so Part I flows naturally into Part II.
  6. Attach the reference list (参考文献): Place your bibliography immediately at the end of Part I in a uniform style throughout (commonly GB/T 7714 sequential numeric format), balancing recent five-year international journals with representative work by scholars in China and mainstream Chinese-language journals.
  • Common pitfalls: Spending four or five pages on broad background before stating a concrete scientific contradiction; claiming flatly that "no prior research exists domestically or abroad" (which alienates specialist referees); or citing too few papers from the last five years while overlooking directly related work by domestic peers—a frequent blind spot for scholars newly relocated from overseas.

Part II: (二)研究内容 (Research Content) — four logical modules

The 2026 official instruction for (二)研究内容 reads: (提纲不做限制,请按照研究工作的自身逻辑撰写。应提炼出特色与创新点、年度研究计划) (Outline unrestricted; write according to the internal logic of your research, distilling distinctive features & innovations and the annual research plan). Although secondary headings are no longer locked, reviewers still evaluate the same core coordinates. Organize Part II into four coherent modules—noting that the 2026 revision moved feasibility analysis and risk mitigation (研究风险的应对措施) into Part III, Subsection 1, so keep Part II focused on design and methods without cross-section repetition.

1. Research objectives, research content, and key scientific questions (研究目标、研究内容与拟解决的关键科学问题)

  • What to write: State research objectives (研究目标) concisely in terms of scientific understanding or methodological breakthroughs achieved by project end; divide research content (研究内容) into 3 to 4 clearly bounded, mutually supporting sub-topics specifying subjects, core variables, and experimental or modeling dimensions; distill key scientific questions (拟解决的关键科学问题) into 1 to 3 mechanistic or methodological bottlenecks that govern overall success (one or two sentences per question plus a brief explanation).
  • Common pitfalls: Listing publication counts or student headcounts as scientific objectives; splitting the work into too many sub-topics for a four-year project; or confusing routine operational chores (sample collection, instrument calibration) with genuine scientific questions.

2. Research design and technical roadmap (研究方案与技术路线)

  • What to write: Map your study design one-to-one against your research content items, detailing experimental protocols, model equations, sample sizes and controls, characterization measurements, statistical analyses, and input–output handoffs across sub-topics. As recommended in the annotated institutional sample, include a visual technical roadmap (技术路线图) linking steps, methods, and stage outputs so reviewers can grasp the architecture at a glance.
  • Common pitfalls: Numbering mismatches between research content and study design; invoking buzzwords like "deep learning" or "multi-omics" without defining features, validation metrics, or controls; or drawing an empty box diagram devoid of concrete methodological detail.

3. Distinctive features and innovations (特色与创新点)

  • What to write: Distill no more than 3 (recommended) innovations across theoretical framing, scientific hypothesis, experimental system, or methodology. Lead each bullet with a concise label of the innovation dimension, then contrast your approach against prior work to show the concrete epistemic gain.
  • Common pitfalls: Diluting your core advance across five or six routine attributes, or overstating the routine application of commercial instruments or off-the-shelf algorithms to a new dataset as an original methodological invention.

4. Annual research plan and expected outcomes (年度研究计划与预期研究成果)

  • What to write: Lay out concrete tasks year by year (for example, 2027.01—2027.12 through 2030.01—2030.12, matching the project period in the online basic information form exactly), and align expected outcomes (预期研究成果) item by item with your research content—covering scientific conclusions, models/datasets, and papers or patents.
  • Common pitfalls: Reusing an earlier draft without updating calendar years so body dates conflict with the basic information form, or scheduling only "literature review and reagent purchasing" in Year 1 without substantive research tasks.

Part III ((三)研究基础) and Part IV ((四)其他需要说明的情况)

(三)研究基础 (Research Foundation) contains 4 locked secondary headings whose Chinese titles and parenthetical text must remain unchanged:

  1. 1. 研究基础与可行性分析 (与本项目相关的研究工作积累和已取得的研究工作成绩,研究风险的应对措施等) (Prior Research Accumulation and Feasibility Analysis, including risk mitigation measures): Present directly relevant prior results and key preliminary data figures, demonstrate feasibility across theory, methodological maturity, and preliminary evidence, and provide concrete fallback plans for 1 to 2 core technical risks. Do not dump unrelated publication lists, and do not omit risk mitigation (研究风险的应对措施).
  2. 2. 工作条件 (包括已具备的实验条件,尚缺少的实验条件和拟解决的途径,包括利用国家实验室、全国重点实验室和部门重点实验室等研究基地的计划与落实情况) (Working Conditions): List available core instruments and platforms factually, and specify concrete arrangements—such as open-fund access at National Laboratories (国家实验室), State Key Laboratories (全国重点实验室), or Ministerial Key Laboratories (部门重点实验室), or partner-institution agreements—for any specialized characterization or computing resources still needed.
  3. 3. 正在承担的与本项目相关的科研项目情况 (Ongoing Related Research Grants): Following the parenthetical instructions, list ongoing related projects held by the applicant and key participants (申请人和主要参与者), specifying funding agency, category, grant number, title, award amount, start/end dates, relationship to the current application, and specific responsibilities—drawing a clear boundary against overlap. Enter 无 (None) if none exist.
  4. 4. 完成国家自然科学基金项目情况 (Completion of Prior NSFC Grant): Detail the completion status, follow-up progress, and logical progression from the preceding completed NSFC project led by the applicant, attaching a 研究工作总结摘要(限 500 字) (work summary abstract capped at 500 Chinese characters) and a detailed output list as required by the parenthetical instructions. Enter 无 (None) if you have no prior completed NSFC project.

(四)其他需要说明的情况 (Other Matters Requiring Explanation) contains 5 locked subsections covering: (1) same-year applications to other NSFC project types by the applicant (applications already declined or not accepted for review need not be listed), (2) institutional affiliation inconsistencies in same-year NSFC applications or participations by senior-titled (高级专业技术职务(职称)) applicants or key participants, (3) institutional affiliation inconsistencies with ongoing NSFC grants held by senior-titled applicants or key participants, (4) same-year applications or participations under different professional titles by the applicant and key participants, and (5) other matters. Cross-check HR and grant records with every key participant; for any subsection that does not apply, preserve the heading and enter 无 (None).

Young Scientists Fund (青年科学基金项目) Categories A, B, and C

Under the 2026 NSFC framework, the Young Scientists Fund (青年科学基金项目) comprises Category C, Category B (formerly the Excellent Young Scientists Fund, 原优青), and Category A (formerly the National Science Fund for Distinguished Young Scholars, 原杰青). Their outlines differ substantially from the General Program:

Program CategoryBody Structure (2026 Official Template)Page / Character GuidanceKey Differences from General Program (面上项目)
Category C (青年科学基金项目(C类))Same 4 parts as General ProgramMain body in principle no more than 30 pagesSole outline wording difference is in Part III, Subsection 3: states 申请人正在承担的…… (ongoing grants of the applicant), omitting 和主要参与者 (and key participants); scope should focus on 2–3 core components
Category B (青年科学基金项目(B类))3 parts: (一)主要学术成绩、创新点及其科学意义; (二)拟开展的研究工作; (三)其他需要说明的情况No overall page ceiling; Part I suggested ≤ 4,000 Chinese characters, Part II suggested ≤ 4,000 Chinese charactersTalent-program structure: Part I highlights innovation and scientific value of prior achievements; Part II details scientific significance, innovation, technical roadmap, and feasibility of planned work
Category A (青年科学基金项目(A类))3 parts with the same top-level headings as Category BNo overall page ceiling; Part I suggested ≤ 5,000 Chinese characters, Part II suggested ≤ 3,000 Chinese charactersPart I emphasizes basic-research achievements over the past 5 years, innovations, research value, and scientific significance (5,000 chars); Part II concisely presents innovative ideas, main research directions, and an initial plan (3,000 chars)

When writing Part I ((一)主要学术成绩、创新点及其科学意义) for Category A or B, avoid stacking chronological paper abstracts; weave representative publications into a coherent academic thread that distinguishes your independent intellectual leadership from collaborator roles. In Part III ((三)其他需要说明的情况), note that the Category A/B wording addresses senior-titled applicants (具有高级专业技术职务(职称)的申请人, omitting 或者主要参与者), and verify every item against the current template downloaded from the NSFC system. For detailed peer-review criteria and age-specific review rules—such as the Northwest A&F University policy briefing noting separate review tracks in 2026 for male applicants under 35 and female applicants under 38—always consult the current year's official NSFC Program Guide (《国家自然科学基金项目指南》).

Generative AI policy in NSFC proposals

In the Notice on Matters Concerning 2026 NSFC Project Applications and Project Completion (《2026 年度国家自然科学基金项目申请与结题等有关事项的通告》, 国科金发计〔2026〕2号; see university-forwarded notice attachment), NSFC sets explicit rules on generative AI:

Applicants who use generative AI technologies to track research trends or collect and organize references when preparing a proposal must manually verify the authenticity and accuracy of AI-generated information and references. Applicants take full responsibility for any generated content used, must disclose such usage comprehensively and truthfully, and must label relevant content in accordance with national regulations including the Measures for Labeling AI-Generated and Synthetic Content (《人工智能生成合成内容标识办法》) and the Guidelines for Responsible Research Conduct (2023) (《负责任研究行为规范指引(2023)》), including but not limited to adding appropriate textual notices at the beginning and end of the text. Proposals directly generated by generative AI are prohibited, and unverified generated content must not be used. (Translated from the official Chinese notice 国科金发计〔2026〕2号.)

Applicants must observe three strict boundaries: (1) never submit a proposal directly generated by AI—your scientific questions and study design must come from your own research; (2) manually verify every literature lead or citation against primary databases and full-text articles; and (3) whenever generative tools assist in organizing materials, disclose that usage truthfully and add labeling notices at appropriate positions at the start and end of the text in accordance with the current NSFC Program Guide.

Organizing and drafting an NSFC proposal in SciVolume

Built around these compliance rules, SciVolume provides preset templates in its grant proposal writer workspace for the NSFC General Program (国自然面上项目), NSFC Young Scientists Fund Category C (国自然青年科学基金项目(C 类)), the National Social Science Fund (国家社科), MOE Humanities and Social Sciences (教育部人文社科), NIH, NSF, ERC, and other funders—organizing a rough draft from your own research materials for you to verify and rewrite paragraph by paragraph:

  1. Preset locked Chinese headings and length limits: Templates lock the official Chinese outline headings and parenthetical instructions, track the 30-page body estimate and 500-character prior-grant summary cap, and let you import existing Word, PDF, or LaTeX drafts mapped by heading.
  2. Draft from your own materials after confirming the plan: Drafting runs only after you enter your key scientific questions, objectives, research content, preliminary data, and working conditions and confirm the full document plan. Missing materials are flagged [待补:…] ([To add: …]) rather than invented; no data, preliminary results, or references are fabricated, and citations draw solely from verified literature.
  3. Free compliance checks and Word export: After rewriting, run reviewer-perspective diagnostics alongside free rule-based checks that do not call a model—verifying required sections, locked headings and parenthetical text, estimated length, 无 entries on inapplicable items, year alignment in the annual plan, innovation point count, technical roadmap references, leftover placeholders, and GB/T 7714 bibliography formatting. Paid plans can generate an editable AI-use disclosure draft from your workspace usage log (placed at the start/end or as an appendix) and export a Word (.docx) file matching the official layout plus a PDF preview, ready to copy into the official Word template downloaded from the NSFC online system (page counts are estimates subject to final layout in the official template; your manuscripts are used only for your project and never to train public AI models).

Newsletter

Join the community

Subscribe to our newsletter for the latest news and updates