As a kind of management strategy of customer centered, Customer Relationship Management (CRM) can retain and obtain customers well. Data mining, as a tool of analyses which can be analyze mass data and mine customer values in CRM. As an important topic in the data mining field—association rules of Data mining, There are some shortcomings in mining association rules via Apriori algorithm or its improved algorithms , for instance, many candidate items generated and more disk I/O needed. After analyzing the disadvantages of the classical Apr- iori algorithm, this paper proposes an improvement of Apriori algorithm based on association rules and apply it in analyzing a large amount of data,for ins- tance,the customers gain, particularly customers distinguish, operation risk and so on. According to the rulers, we can find out people,who have some cha- racters would buy which kinds of insurance. This gave the definitude of opera- tion drumbeating and support enterprise to make an appropriate decision for each kind of customer.