A location planning scheme for electric taxi charging stations based on GPS trajectory data is proposed to address the problem of continuous growth in the number of electric vehicles and insufficient charging facilities to meet user needs. Firstly, the potential charging needs of users are analyzed using the taxi GPS data and demand distribution is extracted. Secondly, a DBSCAN clustering method based on the grid density partitioning is proposed, which reduces the DB index from 0.34 to 0.30 compared to traditional algorithms. The demand is clustered and divided into demand intensive areas, and pre-selected station locations are set. Finally, a set coverage model is developed to achieve site optimization. Using this scheme to simulate Daxing District,Beijing and taxi trajectory data, a reasonable site selection result is obtained, and the results can provide reference for the planning of electric taxi charging stations.