Abstract: In order to find out abnormalities of the electric energy measurement in time and reduce the losses of power enterprises and users,an abnormal diagnosis technology of the electric energy measurement based on measurement automation system is proposed in this paper. The corresponding measurement automation system is designed by using the service-oriented architecture. The system is divided into data collection,information exchange,business processing and other layers;the overall architecture is established,and the system file management,early warning management,real-time meter reading and other functions are set. The mean clustering algorithm is introduced into the system and according to a large amount of historical measurement information,a time series is established to reconstruct the measurement information and fill the missing dat. Using the relative neighborhood information entropy algorithm,combined with the target sample weight,the proportion of abnormal data in the whole domain is calculated,and the outlier is used to diagnose whether the electric energy metering is abnormal. The simulation results show that the proposed technology can effectively fill in the missing measurement data, and has high diagnosis accuracy for equipment abnormalities and user abnormalities.