数字人文为人文研究资料的处理提供了新方法，但在具体实践中却存在重数据分析、轻问题论证与解释的现象，无法有效融入传统人文研究框架。在循证视角下，文章梳理人文研究循证过程，构建SMCI模型描述人文研究问题，从证据检索、证据质量评价、证据综合三个层面开展问题驱动的数字人文循证决策。以南京大屠杀为例，从证据利用检索出发梳理人文研究问题描述中的证据检索、证据质量评价特征及证据综合的具体路径；从证据制作检索出发梳理人文研究问题论证中资料的差异性特征，分析二手证据与问题论证的相关性，并依据论证需求制作一手证据。结果表明，循证视角下的数字人文循证路径，为数字人文开展研究问题有谱系、研究证据有来源、研究结果可解释的数据分析提供理论基础，问题驱动的数字人文循证决策的开展需依据循证范式重构人文研究资料。

Digital humanities provide new methods for the processing of humanities research materials, but in specific practice, there is a phenomenon of overemphasizing data analysis and neglecting problem argumentation and explanation, making it difficult to effectively integrate into the traditional framework of humanities research. From an evidence-based perspective, this paper sorts out the evidence-based process of humanities research, constructs the SMCI model to describe humanities research problems, and carries out problem-driven digital humanities evidence-based decision-making from three levels: evidence retrieval, evidence quality evaluation, and evidence synthesis. Taking the Nanjing Massacre as an example, this paper sorts out the specific paths of evidence retrieval, evidence quality evaluation characteristics, and evidence integration in the description of humanistic research problems from the perspective of evidence utilization retrieval; sorts out the differential characteristics of materials in the argumentation of humanistic research problems from the perspective of evidence creation retrieval, analyzes the relevance between secondary evidence and problem argumentation, and creates primary evidence according to the needs of argumentation. The results show that the evidence-based approach to digital humanities provides a theoretical basis for data analysis with a genealogy of research questions, a source of research evidence, and explainable research results. The implementation of problem-driven digital humanities evidence-based decision-making requires the reconstruction of humanities research materials based on the evidence-based paradigm.