Using the data mining technology to analyze the network learning behavior data can dig out its hidden behavior characteristics, and provide personalized learning resource services for learners . Aiming at the problem that the model applicability of the existing data mining algorithms gene rally was not very high when analyzing of network learning behavior data, this paper proposed a learning resource recommendation algorithm based on behavior sequence analysis. Firstly, this pap er proposed the definition of behavior sequence and its related concepts. Secondly, it proposed the calculation method of behavior sequence similarity. Then, it proposed the collaborative filtering recommendation algorithm based on behavioral sequence similarity to calculate the similarity between learners and to generate the learning resource recommendation list for the learner to be recommended. And then it gave the recommendation algorithm based on learning style, and integrated learners ' learning style characteristics into the recommendation process. Finally, it presented the learning resource recommendation algorithm based on behavior sequence analysis. The proposed algorithm does not limit the pattern of behavior sequence, and has a high applicability, it can be used for further study of the network learning behavior data and provide personalized learning service for learners. [ABSTRACT FROM AUTHOR]