By combined the narrative theory with the needs of archival data processing, the academic concept of data-based event narration of archives is proposed, its core concept, theoretical basis, logical framework and technological roadmap are discussed and verified through a practical case. It is concluded in this paper that narrative is a typical characteristic that distinguishes archives from other literature resources. Data-based event narration of archives refers to the theory, method and technology which under the guidance of historical structuralism methodology, based on binary narrative logic of "frame+object", reconstructs narrative context of core semantic information of archival text or image in the form of structured data sets, constructs computable data model of historical events, and drives intelligent transformation of archival information services. Data-based event narration of archives provides scientific theory and methodology guidance for archival data processing and is expected to have profound impact on the intelligent transformation of archival information system development.