Abstract: In view of the problems of low efficiency,insufficient intelligenization,high labor intensity,distorted on-site monitoring signals and so on in the traditional monitoring mean for mine transformer used by Yangquan Xinjing Mining Coal Industry Co.,Ltd.,based on the analysis of common faults and abnormal characteristics of mine transformers,a mine transformer online monitoring system is designed based on BP neural network theory.The hardware system type selection and software control system design are completed to achieve online centralized monitoring of partial discharge signals of transformers,accurately identifying sound abnormalities,temperature abnormalities,overload faults,and oil leakage faults.After on-site installation and debugging of the 2^# transformer in the downhole of Xinjing Mine,Yangquan City,it shows that the proposed online monitoring system for mine transformers can accurately identify equipment abnormalities.The on-site signals are sent to the upper computer monitoring system through Ethernet,achieving centralized monitoring and remote fault diagnosis of mine transformers.The response time is only 1.83 seconds,with high fault identification accuracy and accurate fault positioning,achieving satisfactory application effects and providing application reference for the later realization of unattended operation and remote operation and maintenance of mine transformers.