Abstract: Aiming at the state of subway train components interacts due to coupling，a design method of metro train state coupling analysis system based on pattern recognition is proposed. The method extracts the anomalous features through wheel-rail coupling analysis，bow-net coupling analysis and component state feature recognition technology，and acquires the faults caused abnormal features. Combined with the train of a subway company，the service security system is designed and developed. The research results show that the method can obtain the anomaly features information through the condition pattern recognition and feature coupling analysis，and effectively reduce the interference by obtaining the anomalous features to analyze the fault causes，so as to lay the foundation for rapid positioning and accurate maintenance.