This work is in the context of wind energy, which is the energy source that grows more in the Brazilian energy matrix, according to the Energy Research Company (EPE), with projections that this growth will continue. Thus, the main motivation of this work is the fact that developing and implementing increasingly precise forecasting methods for the key variables in the production of wind energy in a wind turbine, such as wind speed, is of crucial importance for planning of the national electric system operation. Therefore, the main objective of this work is to adapt and apply a time series forecasting methodology in a database formed by wind speed measurements. The methodology is built from the exploratory analysis of data, which can be observed important features such as stationary mean and a complex seasonal structure, which involves a daily cycle and monthly seasonality. Thus, it was adapted an exponential smoothing model that incorporates multiple cycles, Monte Carlo simulation and decomposition of the series through the TBATS method, to make forecasts. As results and conclusions, it is possible to observe that model adapted was adequate to address the proposed issue, compared with the forecast models established in the literature, resulting in an increase in the accuracy of forecasts made.