In order to improve the precision of the load forecasting, some related factors are chosen in the process of establishing a load forecasting model. However, some factors that are not very important are taken into consideration, which increases the difficulty of establishing a load forecasting model. Even sometimes it is difficult to explain the model established. In order to avoid this kind of situation, the RE(ridge estimate) method is firstly used to select the independent variable set, and to exclude those factors which are not significant. Then, the PLSR(partial least squares regression) method is applied to establish the model based on the rest of the independent variables set. Through the practical example, it is shown that the method proposed in this paper is correct and effective.