Existing electric vehicle charging station (EVCS) planning and design studies give less attention to the multi-type charging mode containing high power charging. Thus a two-stage siting and capacity determination method is proposed for the charging station planning problem of multi-power level charging demand. First, a charging load spatial and temporal distribution prediction model considering dynamic transfer planning is constructed based on real-time road conditions. Secondly, the probability of users choosing different charging stations is analyzed to determine the service range of each charging station, considering the scale of the charging station and user charging accessibility. Then, aiming to minimize the annual total investment cost of charging stations, collaborative planning of the distributed energy storage system (DESS) and multi-power level charging facilities is conducted. Finally, the effectiveness of the method is verified by taking the topology of IEEE33 node coupling between the road and distribution networks as an example. The simulation results show that the method can enable high-power charging facilities to reasonably access the grid, improving the economic efficiency of charging station planning while ensuring the security of the distribution grid.