Abstract: Aiming at the uncertainty of electric vehicle (EV) grid connection time and charging time, we proposed a robust optimal scheduling model for grid-connected electric vehicle cluster based on support vector clustering (SVC). Taking the charging and discharging power of EV as the decision variable and the minimum charging cost of the user as the objective function, we established a cluster EV scheduling model. Using the EV historical charging data, and taking the smallest hypersphere containing all sample data as the shape of the uncertain set, and the generalized histogram cross-kernel as the kernel function, we calculated the uncertain sets of EV network access time and charging time parameters, and established. A rod optimization scheduling model of cluster EV based on SVC. The analysis of the final example shows that the method proposed in this paper can describe the uncertainty parameters of EV charging more accurately, and the model can quickly respond to the time-of-use electricity price while ensuring economy, and has good practicability.