Abstract: 【Purpose/significance】This paper extracts topics and analyzes evolution trend in information resource management in China based on the deep learning topic modeling method, and provides reference for related research and practice.【Method/process】 The literature data of CNKI were collected, and the topic identification and evolution trend analysis of domestic information resource management journal papers in recent five years were carried out by using BERTopic model and dynamic topic model.【Result/conclu⁃ sion】The research on information resource management includes 50 main topics, among which there are more papers related to library reading service, archives and digital humanities, online medical and health, public cultural service, intelligence work and intelligence analysis. The 50 topics are grouped into 10 research directions, including library management and service, bibliometric analysis, user information behavior, archive management and digital humanities, open data management, competitive intelligence and think tanks, network public opinion research, smart library construction, information security and rights, and emergency services. It is found that the themes of disruptive technology identification, online medical and health care, video user public opinion, personal privacy rights protection, cross-border data flow and data sovereignty have obvious growth trends; The trend of competitive intelligence and think tank is on the rise, especially the trend of intelligence work and intelligence analysis. The popularity of archival science related topics shows a steady trend; The direction of library management and service decreased obviously, but the theme of public cultural service in⁃ creased slightly. Open data management topics decreased slightly; The popularity of online rumors, public health emergencies and li⁃ brary epidemic services continues to decline; Emerging technologies such as ChatGPT, artificial intelligence, and metaverse appear in multiple directions with prominent cutting-edge features. On the whole, the change of research direction and theme shows the charac⁃ teristics of advancing with The Times.【Innovation/limitation】The BERTopic model is used to analyze the overall development trend in the field of IRM in China. The model will be used to explore the correlation and changes of different topics in a specific research di⁃ rection. And the topic recognition effect difference of different modular method combinations of the model will be compared to find the method combination with the best effect.