Under the circumstances of development for scaled electric vehicles (EVs), it is of great significance to rationally control and optimize the EV charging load and carry out the structure planning of the active distribution network. This paper establishes a bilevel model for structure planning of active distribution networks based on the regulation of the EV charging load. Firstly, taking the uncertainties of distributed generation and load into consideration, the main process framework for EV charging load forecasting is constructed, and then the dispatching method of the charging load is described and the mathematical model of structure planning for active distribution networks is build up. In such a mathematical model, the planning layer aims at the lowest comprehensive economic cost for kilowatthour of electricity and the operational layer is targeted at minimizing the load curve variance. Secondly, in the process of model solving, the Prim algorithm is utilized to improve the population generation process of particle swarm optimization (PSO) and constraints are processed by a penalty function so as to improve the efficiency and precision of algorithm. Finally, the rationality and validity of the model are verified by an actual distribution network.