To alleviate the adverse effect of large-scale electric vehicles (EVs) random charging, an EV hierarchical charging method considering multiple modes coordination is proposed in this paper. This proposed method avoids the large-scale charging load centralized in a certain period by coordinating diverse charging modes. Considering the load characteristics such as the output of renewable power generation, the power load curves of an area in Inner Mongolia are clustered into five typical power load curves based on the K-means clustering algorithm and elbow method. According to EV charging and battery swapping (BS) modes’ characteristics, the charging station models and EV hierarchical power exchange model are established. At the upper level, the users’ requirements and charging costs are considered, and EVs are matched with the charging mode based on the particle swarm optimization algorithm. At the lower level, the electric price and charging station operation conditions are considered, and the charging schemes in the charging/BS station are optimized based on the optimization toolbox in the Matlab software platform. Extensive case studies are presented to validate the effectiveness of the proposed method, where a large number of EVs charge continuously with cost efficiency. Moreover, through the charging power and the typical power load curves' respective superposition, the analysis demonstrates that only electric price guides EV charging load may exacerbate the valley-to-peak and instability of the local power load curve.