Abstract: Coal mine safety management system is an important means to maintain the order of coal mine safety production. Good rules combined with the implementation of the system can play the effectiveness of the system, but high implementation intensity means the increase of management cost. This paper discusses the impact of the characteristics of system implementation on the effectiveness of the system, so as to provide a reference to improve the effectiveness of the system under resource constraints for coal mines. Based on the attitude theory and reinforcement theory of psychology and behavior discipline, this paper puts forward the basic viewpoint that the subject of coal mine safety management system has a specific tendency to the cognition of the system rules system, and then forms the system compliance will, and the system implementation system regulates the relationship between the system compliance will and the system behavior. On this basis, under a specific level of system compliance will machine learning method is used to simulate the impact of system implementation characteristics on the emergence of system compliance behavior, and to help to reveal the appropriate intensity of system implementation.