This paper presents a charging infrastructure planning method considering multi-dimensional charging behavior constraints of EVs. Firstly, the whole area to be planned is divided into grids by trunk road network or municipal units, and the number of vehicles in each grid is determined based on the predicted number of EVs in the target year. On this basis, a matrix is constructed from the dimension of a single vehicle to describe the full-duration charging behavior, and a three-dimensional matrix is constructed from the dimension of the grid to describe the full-duration charging behavior of the vehicle belonging to the grid. Then, taking the above matrix as the optimization variable, a multivariate nonlinear optimization model with the minimum social cost as the optimization objective is constructed. Through the simulation and analysis of the proposed planning method under the simulation of real county scenarios, and the introduction of electricity price mechanism, the rationality and economy of the proposed method are verified.