This paper aim to analyses wind prediction over the Alagoas State (Brazil) using the WRF model. A 1-yr (August/2007 to July/2008) anemometer collected data is used to analyze wind pattern and the model predictions were statistically validated using stations located in the three State mesoregions: Backlands (Sertão), Rural (Agreste) and Coast (Litoral). The results showed that quality of WRF model prediction proved to be quite satisfactory especially in the interior of the State during the dry season. The wind predictions in the rainy season for coastal zones showed bias of 1.77 ms-1 and RMSE of 3.61 m.s-1, while in the dry season these ratios were 0.98 m.s-1 and 2.99 m.s-1, respectively. In the interior of the State, these indicators reached bias of -0.2 m.s-1 and RMSE of 2.75 m.s-1 for dry months, and bias of -0.1 m.s-1 and RMSE of 2.63 m.s-1 for rainy season. Some aspects of the WRF model must be further tested and analyzed to improve the prediction during the rainy period, especially the parameterizations of cumulus and clouds microphysics. The obtained statistical indexes were equivalent or better, in some cases, if compared to other similar studies, indicating that the WRF is a good tool for wind forecasting.