Deeply research metering data with the big data analysis platform for the distribution network operation is one of the key research directions in the power grid currently. Power outage event completion method based on SVM algorithm to solve accurate statistical problems for MV distribution network was researched in this paper. Firstly, based on summarizing the five types of power outages in the MV distribution network, the SVM completion method is mainly researched. And the power outage event completion idea based on SVM and a full-process of SVM completion model construction method for the five types of power outage events are given. Besides, the completion process of distribution network model construction, vector machine SVM construction, vector machine SVM solution and judgment of fault type are proposed. Then, the business relationship framework among the complementary modules and other related systems, the data analysis architecture based on the big data platform are designed from the perspective of engineering application. Finally, take the application of Anhui Huangshan and other three cities as an example, it verifies that the completion method can greatly improve the timeliness and accuracy of power failure event collection.