The invention discloses a wind power forecasting method based on a wind district wind belt wind speed rule. The wind power forecasting method comprises the following steps of establishing a neural network; training a model of the established neural network by applying the historical data such as wind speed, wind direction, temperature, air pressure and humidity and corresponding historical output power in a wind power station, namely obtaining an initial wind power forecasting model; inputting the data of wind speed, wind direction, temperature, air pressure and humidity to the wind power forecasting model so as to obtain forecasted results. The purposes of accurately forecasting the power of the wind power station and thus guaranteeing the power grid safety are achieved.