Aiming at the problems of low execution efficiency, excessive I/O burden and large amount of calculation of traditional association analysis Apriori algorithm, an improved Apriori algorithm was proposed, which reduced the computational complexity of the candidate item set by reducing the number of database scans, and improved the execution efficiency of the algorithm by converting the transaction item set into a row vector and using the “and” operation. The effectiveness and efficiency of the improved Apriori algorithm were verified by using the data set of students’ behavior. At the same time, in order to meet the requirements of the algorithm for sample data, the original data were cleaned and discretized in the process of sample data processing, and the rules for discretization of sample data of the analysis object were defined. The performance of improved Apriori algorithm and classical Apriori algorithm was analyzed and compared through experiments. The results show that the improved Apriori algorithm maintains the effectiveness of mining association rules for actual analysis objects, and has higher execution efficiency.