The thesis consists in the creation of a wind power production model of mesoscale areas, like states. Inputs of the model are the production history of a geographical area, weather data, total installed capacity and possibly power capacity distribution. In case this last datum is not available, it is still possible to indirectly estimate it and train the model through an inverse mathematical problem. Due to the large amount of calculations, different optimization schemes are compared, in order to achieve a good solution in the most efficient way.n using climate reanalysis