Photovoltaic systems’ faults can cause a power reduction, the decrease in the useful life and irreversible damages on the photovoltaic cells. This paper presents a strategy able to diagnosticate the photovoltaic array fault according to the Brazilian regulatory standard NBR 16274. Therefore, the fault diagnostic is based on the PV array I-V curve which is used to calculate the inputs of a multilayer perceptron artificial neural network (ANN). The results shown in this paper demonstrate the ANN with a compact structure and good accuracy. Moreover, it shows the ANN can classify all faults of the NBR 16274 with 99.3% of accuracy https://doi.org/10.53316/sepoc2022.025