Abstract: In order to reduce the high line loss rate of power supply service caused by abnormal data in the station area, the SVM technology is introduced to design the treatment method of abnormal line loss data in station area. Sensors and other related equipment are used to collect operation feedback data of power equipment in the station area. The clustering center of the abnormal line loss data in the station area are set to identify and cluster the abnormal line loss data in the station area. The VisuShrink technology is used to determine the threshold of data processing. Using wavelet threshold method and refering to wavelet coefficients abnormal signals are reconstructed to eliminate background noise of signals. The SVM support vector machine is introduced, by setting the training sample set and using the machine learning algorithm in SVM, the regression function is established, further through assisting the POS parameter optimization conditions in the support vector machine and controlling the particle spacing of spatial particles, the line loss abnormal data in the station area is governed by reconstructing the spatial abnormal data. Taking a large power enterprise in a certain area as an example, a comparative experiment is designed. The results of the experiment prove that the designed governance method has a good effect in practical application, and can stabilize the line loss rate in the station area below 5%.