Based on the developed and used management system of fitness club, puts forward the application of C4.5 decision tree classification algorithm for the fitting-record data mining. By this method finds out the members' age, gender, membership card types and the participating time rules in their club current validity period, and extracts fitness participate member features at a specific period of time. The experimental results show that the classification rules applying to the member management system can help health club managers make targeted marketing scheme.