Electric vehicles connected to the grid exhibits strong uncertainty in time and space. Dumb charging of large-scale electric vehicles might have adverse impacts on distribution system by causing much high peak load exceeding the supply limits of devices. This paper proposes a model for smart charging control of electric vehicles, which takes smoothing the daily load profile as the objective function and fully accounts the EV owner’s requirement. An adaptive genetic algorithm is applied for solving the model. Using the IEEE 33-bus case as the test systems, scenarios of EVs integration are simulated by Monte Carlo stochastic methods. Smart charging strategy is obtained using the proposed model and method. By comparing with the load profile under dumb charging, the validity of the proposed model is proved.