The accuracy of load forecasting of power system is the guarantee of the safe, stable and efficient operation of the power grids. Least squares support vector machine (LSSVM) is widely used on the load forecasting of power system, but this method has many shortcomings in dealing with uncertainty problems. In order to improve the accuracy of selecting the parameters of the kernel function, to deal with uncertainty factors and improve the accuracy of short-term load forecasting, this paper proposes a new model which is combined by the cloud model, particle swarm optimization (PSO) and LSSVM. First of all, through analyzing uncertainty of each influence factor, it uses the models of Cloud-LSSVM and PSO-LSSVM separately to predict the impact factor according to the uncertainty, then it achieves the final forecast through the weighted combination model. At last, the simulation of experiment proves that the new model can achieve better load forecast of power system by dealing with the uncertainty factors.