The invention relates to an intelligent prediction method for wind power output. The method comprises steps: step S1, collecting wind power output power and meterological index data; step S2, performing moving smoothing processing on the collected data; step S3, establishing a support vector regression model, using the wind power output power which is performed by moving smoothing processing as an output variable, using the meterological index data which is performed by moving smoothing processing as an input variable, training the support vector regression model; step S4, optimizing parameters of the support vector regression model, training the model by using the input variable and the output variable in the step S3 again; step S5, obtaining to-be-predicted meterological data, after moving smoothing processing, inputting the support vector regression model which is optimized and is trained again, to obtain a wind power output power prediction value. The method can realize accurate prediction of short-term wind electricity output power.