Abstract: 【Purpose/significance】To promote the efficient organization and inheritance of historical resources related to the AntiJapanese War, enhance academic research on the war, and improve public awareness of the war, this paper proposes an event-driven method for constructing a knowledge graph of the Anti-Japanese War.【Method/process】This study first collected and organized 1, 022 Anti-Japanese War events from historical resources. Based on an event-driven perspective, the study identified the temporal, causal, and containment relationships between events, constructing the corresponding triples. Then, entity extraction and relation rec⁃ ognition were conducted using large language models. Finally, the constructed knowledge graph was imported into the Neo4j database for visualization and storage to support applications such as intelligent question answering and historical analysis.【Result/conclusion】 The research indicates that the event-driven knowledge graph of the Anti-Japanese War can systematically organize historical re⁃ sources, breaking the limitations of traditional historical narration. It provides multi-dimensional analysis tools for academic research and offers an interactive historical learning experience for the public, contributing to the inheritance and promotion of the AntiJapanese War spirit.【Innovation/limitation】This study proposes an event-driven method for constructing a knowledge graph of the Anti-Japanese War and combines large language models for entity and relation extraction, facilitating the systematic organization and deep exploration of historical and cultural resources.