The article proposes a methodology applicable to any photovoltaic (PV) plant to obtain an approximation of the monthly production of solar array power. The analysis was carried out in seven systems, of different technologies and capacities, connected to the microgrid of the Center for the Development of Renewable Energies (CEDER) belonging to the Center for Energy, Environmental and Technological Research (CIEMAT) in Soria, Spain. The proposal simulates radiation by combining and crossing two Gamma probability distributions, representing the days with the best and worst solar resources, respectively. As a result, a matrix was created with 12 variables that define the monthly behavior of the radiation. On the other hand, the granularity of the PV generation was homogenized to know it at any moment through polynomial functions. Once both characterizations were known, it was possible to predict the monthly power of each PV array. The methodology has been validated with the measurement approximation index, developed in the text, and with specialized software. The results presented will help in the dimensioning of a backup model and will collaborate in the adequate management of energy.