For an effective planning and operation of the Brazilian electricity sector, reliable time series of renewable energy generation is required, which are of great importance for the provisioning of operating needs and costs. These time series usually come from measured data from generations that occurred in different locations. However, in the absence of such series, reanalysis bases have emerged as a promising option. In the case of wind generation, it is possible to obtain, through the reanalysis bases, historical series of wind speeds with long periods and available for almost the entire terrestrial surface. One of the disadvantages is the arrangement of data in a grid and with low spatial resolution, not covering all points on Earth. As a way to overcome these limitations and bring wind speed series closer to the most probable values in the real world, the present study uses interpolation, extrapolation and bias correction strategies, as suggested by the specialized literature. Then, the obtained series are compared with historical series recorded in Brazilian territory to assess their applicability in Brazil. The results achieved indicate that the wind speed series extracted from MERRA-2 have relevant representation of the historical series measured in the locations under study.