power load model plays an important role in power system planning, design, operation and control. however, because of some complex characteristics of power load, such as randomness, time-varying, nonlinear, the problem of load model has not been solved properly. The main object of this study is a new intelligent optimization search algorithm--- Gene Expression Programming (GEP), it doesn’t rely on worker’s too much experience, not to mention the need to specify the objective function structure in advance. By using the algorithm of GEP, program can automatically help us find hidden relationships between data inside, thus the automatic modeling of function comes true, and then we can find the optimal load model.