This paper proposes forecasting aerodynamic power of wind in order to improve different supervision algorithms of wind farm based on artificial neural network ANN. The study described in this paper develops a simple and robust algorithm that describes short-term wind power forecasting. As wind energy varies during day time depending on the wind speed hitting the generator blades, If an accurate prediction of the wind speed for the following hour can be evaluated, the total amount of active power that can be produced by each generator on a wind farm can be determined and therefore, the amount of energy that could be sold during the next hour would be known too. A set of recent wind speed measurement samples from meteorological stations at ADRAR located in the south of Algeria, are used to train and test the data set. The performance of the proposed algorithm is verified by using MATLAB software. The result obtained has given rather promising results in view of the very small mean absolute percentage error (MAPE).