In order to reduce the blindness of charging station construction and improve the overall satisfaction of charging station construction and operation enterprises and electric vehicle users, this paper researches the location of electric vehicle charging station. First, the number of electric vehicles in the future is predicted by the grey model. The electric vehicle charging demand is calculated based on the road network, traffic flow and user charging behavior characteristics. Then the charging station service range is quickly divided by the Voronoi diagram. And a mathematical model is established based on the minimum total social cost. Finally, Improved particle swarm optimization algorithm and Voronoi graph are combined to solve the problem. The results of the example show that the algorithm can efficiently and accurately obtain the site selection scheme with the minimum social total cost, which is of great significance to urban development planning.