Most distribution network accidents in coastal areas are caused by natural disasters, among which the typhoon is the most serious. Although a typhoon is a low probability weather event, once it appears, it will have a serious impact. Making a scientific and effective optimization strategy for emergency repair and recovery is conducive to improving the speed of power restoration after a disaster. This paper presents a two-stage optimization strategy for emergency repair and recovery of distribution network. In the first stage, the fault probability of the distribution network line is calculated from the multi-source heterogeneous data monitored in a typhoon disaster. Then a location model of a temporary allocation center for emergency repair of the distribution network is established with the goal of minimizing the loss of load caused by the fault line waiting for emergency repair. In the second stage, considering the uncertainty of distribution network fault after a typhoon disaster, Monte Carlo scenario generation and synchronous back generation scenario reduction are used to generate a post-disaster distribution network fault scenario. Based on the idea of cooperation between mobile distributed power supply and emergency repair personnel, an improved genetic algorithm is used to analyze the optimization model of post-disaster distribution network fault emergency repair and recovery. Finally, the IEEE 33 bus system is used to verify the scientific nature and validity of the proposed method. This work is supported by the Science and Technology Project of China Southern Power Grid Co., Ltd. (No. GDKJXM20198441 (036100KK52190053) and No. GDKJXM20210044 (036100KK52210047)).