Abstract: The irregular missing of user electricity data in the collection and transmission of low-voltage station can lead to errors in the subsequent application analysis of the station area. In order to ensure the integrity of electricity data in the low-voltage station,this paper proposes a multi-user missing electricity data complementation method based on the residual U-Net（RU-Net）. First,according to the characteristics of missing multi-user electricity data in low-voltage stations, the electricity data are formed into a spatio-temporal tensor data format that can be operated by one-dimensional convolution. Then, the coding and decoding capabilities of RU-Net are used to reconstruct the missing data, and the network structure is optimized by introducing residual learning and batch normalization layers. Finally, based on the proposed method, two cases of random missing and consecutive missing user electricity data in some stations are complemented. The results show that the method can complement the electricity data of a station with random missing rate not more than 40% and consecutive missing rate not more than 2 days. The proposed method has a certain degree of improvement in precision compared with the traditional method.