In order to improve the accuracy of wind power short-term forecast, a kind of new combination forecasting model is proposed combining the maximum-minimum approach degree with induced ordered weighted operator. According to different induced ordered weighted operators, different combination forecasting models can be constructed, namely IOWA combination model, IOWHA combination model and IOWGA combination model. Because real data are unknown in the forecast period and induced value of every single forecasting method can’t be calculated in advance, this method can not be used to predict directly. Forecasting value of the highest accuracy combination forecasting model selected from different combination models built by single forecasting models instead of real data can solve the induced value calculation. Simulation results of two different wind power farms show that compared with some single forecasting models and another combination models, the forecast precision of IOWGA combination model is lower and predictive effect isn’t improved; forecasting error indexs of IOWA combination model and IOWHA combination model are all lower and forecast precision of those is higher, moreover, predictive effect of IOWHA combination model is better than IOWA combination model, therefore, taking forecast value of IOWHA combination model as final prediction result can effectively improve the prediction accuracy of wind power.