Abstract: To reduce the loss caused by power outages,this paper proposes a two-stage intelligent allocation technology for repair crews under typhoon disasters. Firstly,meteorological factors,geographical factors and power grid factors are considered to predict the number of users encountering the power outages on the grid scale of 1 km×1 km using the random forest algorithm. Then the damage degree of each grid is graded,and the non-dominated sorting genetic algorithm is used to optimize the repair path according to the damage degree.Finally,the effectiveness of the allocation technology is verified by an example of a county in Guangdong. The results show that the forecasting accuracy of the number of users encountering the power outages can reach 88.48%,and the optimization of the repair path can help to choose the optimal allocation strategies of the repair for power grid companies,effectively coping with the typhoon disasters.