Abstract
This article compares the predictive performance of ARIMA and artificial neural networks. Production forecasts of wind power to produce electricity in the field of "Kabertne" (70 km north of the State capital of the Adrar, in southern Algeria), the results show that ANN mode better predict than ARIMA model, according to the criteria of the MAE, RMSE, and MAD. overall ANN are the best model and can be used as an alternative method for the prediction of wind energy production. The actual empirical results with data indicate that the proposed model can be an effective way to improve the accuracy of forecasting by artificial neural networks. Therefore. it can be used as an alternative model appropriate for the operational forecasting, especially when forecasts more precision is needed.