As environmental pollution and energy crisis become more and more serious, electric vehicles have attracted widespread attention due to their advantages of low emission and low noise. But a flood of electric vehicles plugged into the grid will have a huge impact on the power system. Especially in the living area, EV users are generally used to enter the community to charge their cars immediately after work, and this period is the peak period of household electricity consumption. Large-scale uncontrolled charging behavior will cause security risks to the power grid and affect the normal operation of the power grid. On account of this phenomenon, a two-level programming model aiming at ‘peak cut’ for time-of-use price of electric vehicles in residential areas was established. The lower was minimization model of charging cost for lower users, getting the charging load. The upper was load time distribution equilibrium model of upper power grid, obtaining time-of-use price of fast and slow charging power. It could change the charging behavior of residents by time-of-use price. Finally, the calculation example showed that the power supply stability of the grid was improved by 57.66%.