With the scaled development of electric vehicle, it is urgent to make reasonable site selection planning for charging stations to meet the actual needs. Electric vehicle traffic trajectory data are modeled and analyzed, and spectral clustering method is adopted to realize rapid clustering of electric vehicle driving data. According to the results of rapid clustering, the optimal location planning of charging stations is realized with the goal of minimizing the cost of construction economic operation and maintenance after the location of charging stations. The result of the calculation example shows that the charging station established can satisfy the user’s convenience and minimize the annual economic cost.