Objective: Detect the hot spots present in the photovoltaic modules, to know what type of maintenance they need to achieve good energy performance.

Methods: The bibliographic review method, the inductive-deductive and the experimental method were applied, through visual inspection and infrared thermography to determine the number of panels that are affected, using the temperature variation technique.

Results: It was obtained that 7.41% need corrective maintenance, 11.11% preventive maintenance and 81.48% predictive maintenance, this maintenance must be carried out with the objective that the photovoltaic generator does not lose its optimal operability during its work process and decrease its performance, this prediction method is done automatically, the procedure allows you to organize the detection and classification process more effectively.

Conclusions: Thermography is a passive, non-contact measurement method. Thermal images show the temperature distribution on the surface of the object. Thermal imaging technology has become one of the most valuable diagnostic tools for predictive maintenance by detecting anomalies that are not normally visible to the naked eye. These can prevent costly system failures before they occur, thereby achieving that reduces the occurrence of breakdowns, increases the overall performance factor, improves the profitability conditions of the project, reduces costs associated with its operation.