Forest fires are global phenomena that pose an accelerating threat to ecosystems, affect the population life quality and contribute to climate change. The mapping of fire susceptibility provides proper direction for mitigating measures for these events. However, predicting their occurrence and scope is complicated since many of their causes are related to human practices and climatological variations. To predict fire occurrences, this study applies a fuzzy inference system methodology implemented in R software and using triangular and trapezoidal functions that comprise four input parameters (temperature, rainfall, distance from highways, and land use and occupation) obtained from remote sensing data and processed through GIS environment. The fuzzy system classified 63.27% of the study area as having high and very high fire susceptibility. The high density of fire occurrences in these classes shows the high precision of the proposed model, which was confirmed by the area under the curve (AUC) value of 0.879. The application of the fuzzy system using two extreme climate events (rainy summer and dry summer) showed that the model is highly responsive to temperature and rainfall variations, which was verified by the sensitivity analysis. The results obtained with the system can assist in decision-making for appropriate firefighting actions in the region.