In this paper, combining artificial neural network (ANN) prediction model with gray prediction model (GM) effectively as an optimal combination forecasting technique is proposed. It can reduce the prediction error when it is applied in wind power generation capacity forecasting. Taking the factors affecting the wind power generation capacity into account are wind velocity, wind direction, temperature, and the wind power generation amount in previous years and so on. Combining with the advantages of both ANN and GM model, using the optimal combination of the forecasting techniques can improve the prediction accuracy and reduce the prediction error. From the result, the optimal combination of the forecasting techniques error is less than a single gray prediction and neural network prediction. It has research value for the wind power generation foresting in the future.