The detection of fires by orbital data can be done using specific techniques, such as the Normalized Burn Ratio (NBR), which assists in the identification of burned areas, since the affected ones present characteristic spectral responses. More sophisticated studies use satellite data to map different types of fire severity present within burned areas. The objective of this work is to apply the indices NBR and dNBR to identify burnings and their degree of severity. Data from the MUX - IRS (CBERS - 4) and OLI (LANDSAT - 8) sensors were used, performing a comparative analysis between the results found on different sensors. The obtained data were preprocessed and properly prepared for the calculation of the indices and the results for the different satellite images used were confronted visually and statistically. We conclude that the use of NBR and dNBR is an important tool for the burned area classification in the study area. It was also observed that the results obtained with the Landsat-8 image presented better efficiency in relation to the CBERS-4 data for the identification of burnings and their severity.