The development of computer networks provides learners with new educational platforms for knowledge acquisition and skill learning. In this study, a learning behavior analysis model based on distance education platform is formed, and clustering analysis, lagged sequence analysis and association rule mining are used to visually analyze learners’ online learning behaviors and the strength of association between behaviors. In this paper, 76 effective learner behavior sequences are extracted based on the distance education platform, among which the frequency of watching learning videos is the highest, at 1264 times. According to the eigenvalues of different behaviors, learners are divided into 5 clusters. The behaviors of learners in different clusters changed before, during, and after the course. The three indicators of video watching, submitting assignments, and chapter testing have the greatest influence on learning performance, with correlation coefficients of 0.689, 0.616, and 0.561, respectively. The different behaviors of learners are interdependent, and there are correlation rules for different strengths.