This paper proposes an algorithm for electric power business system function optimization. Firstly we extract log data from the Web server and the client user. Then, we preprocess the user log dataset, and convert transaction dataset into a sequence data. Finally, we use the improved Apriori-based algorithm to find tight coupling function modules, and make optimization combination between the functions to improve the working efficiency of the business. Experiments show that the method is effective in revealing the coupling of the business function module.