The influencing factors of the comprehensive monitoring cycle of the centralized monitoring substation are analyzed, including the voltage level of the substation, the frequency of centralized monitoring defects, equipment operation situation, the importance of the substation, the operation environment and the average load of the equipments. Those information can be obtained by constructing the knowledge map of centralized monitoring. On this basis, a comprehensive monitoring cycle classification method based on competitive neural network is proposed, and machine learning method is used to achieve a scientific and reasonable classification of comprehensive monitoring cycle of the substations. The simulation results show that the method is feasible and effective.