Abstract: 【Purpose/significance】To promote the innovative development of MOOC and online teaching, and to build high-quality on⁃ line open courses, it is necessary to explores and analyzes the needs of online course users.【Method/process】With user comments of online courses as data source, firstly, topic extraction of needs for online course users is carried out based on BERTopic. After that, the extraction results are visualized and analyzed to determine the elements of user needs. Then, the KANO model is used to classify user demands, and finally, corresponding practical insights and suggestions are proposed.【Result/conclusion】The results indicate that the video, audio, and picture quality, as well as teaching content of online open courses, are the basic needs of users and must be fully met by course builders; Users also have high-level demands for course design, teacher explanations, teaching assistant teams, course re⁃ sources, as well as subtitle and bullet screen functions. Secondly, the various needs of online open course users are in a dynamic evolu⁃ tion, with some charismatic needs gradually shifting towards expectant needs, while some previously identified factors that can affect user satisfaction are being ignored by users. Therefore, the construction of online course should continuously tap into the implicit needs of users and enhance the competitive advantage of the curriculum while ensuring that the basic and expected needs of users are met【. Innovation/limitation】This study presents an innovative method for mining online course users, which integrates BERTopic and KANO models. However, only Python online open courses were selected for empirical analysis, and whether they are applicable to on⁃ line courses in other fields needs further testing.