Accurate fault location of distribution network is of great significance to improve the reliability of power supply and reduce the loss caused by continuous power outage. The application of micro-phasor measurement units(μPMU) in distribution network provides more information for accurate fault location. Therefore, a distribution network single-phase grounding fault location method based on the extreme gradient boosting (XGBoost) of μPMU information and the support vector unit model based on genetic algorithm (SVM-GA) is proposed. Firstly, the fault section is judged by the zero-sequence current direction provided by μPMUS. Then, based on the positive sequence vector of terminal voltage and current and the feature set of actual fault distance, the combined algorithm distance model is trained. Finally, the combined algorithm fault locator is used to locate the fault of the verification set. Through Matlab/Simulink, it is proved that this method can effectively locate the fault and can withstand the influence of transition resistance. Compared with the commonly used method, the positioning accuracy of the combined positioning method is higher.