Abstract: In view of the problems of low detection accuracy and long detection time in the detection of abnormal electricity consumption state of intelligent electric energy meter, a detection method of abnormal electricity consumption state of intelligent electric energy meter based on multi-source data fusion is proposed. According to the extracted data abnormal characteristics, the multi-source abnormal data of intelligent electric energy meter are obtained. The obtained abnormal data are standardized by using the feature normalization method and the inverse spectrum mean subtraction method. The multi-source data fusion is used to recognize and detect the abnormal electricity consumption state after standardization processing by using the gate control loop unit (GRU). It is proved by experiments that the designed method has IoU above 0.8, and the detection time is not more than 1 s, which can achieve the precise real-time detection of the abnormal electricity consumption state of intelligent electric energy meter.