The access of a large number of electric vehicle loads will change the distribution of power flow in the distribution network and threaten the stability of the power grid. Based on the semi-invariant and Gram-Charlier series, this paper calculates the dynamic probability flow of the distribution network considering the charging load of electric vehicles. On this basis, a comprehensive index evaluating the influence of the charging station on the power flow of the power distribution system is defined based on the risk indicator of branch overload and bus voltage violation, and a method for determining the optimal electrical access point of the charging stations is proposed. The simulation results of IEEE33 node power distribution system show that using this method to select the electrical access point of the charging stations can reduce the risk of branch overload and bus voltage violation. It can be used to guide the charging behavior of electric vehicles through an optimized layout of the charging facility. This work is supported by National Key Research and Development Program of China (No. 2018YFB0904604), The Technical Standards of Shanghai Science and Technology Commission (No. 18DZ2205700) and Shanghai Minhang District Industry University Research Cooperation Project (No. 2017MH271).