This study had aimed to observe changes in land use and occupation and to associate with the surface temperature (ts) of the municipality of Pombal-PB for the period from 1986 to 2016. In this research, bands 3, 4 and 5 were used. TM / Landsat 5 sensor and OLI / Landsat 8 bands 4, 5 and 6 to analyze changes in land use and occupation. To determine the TS, two multispectral images (thermal infrared bands) were used, through TIRS sensors (thermal infrared sensor), referring to bands 6 and 10, Landsat 5 and 8, respectively, in which they have resolutions of 30 meters, for years 1986, 1996, 2006 and 2016. To determine the classes of land use and occupation, the maximum likelihood method (maxver) was applied. To measure the compositions of the classes determined by maxver, the statistical technique, Kappa index (k), was used. The results showed that in the surveyed time scale (1986 to 2016) that the classes of land use and occupation that showed growth were anthropized area (13.71%) and exposed soil (12.01%), in contrast, the Caatinga class thinning showed a reduction (48.76%). The results of the Kappa index indicated an excellent degree of performance for the four periods analyzed. The results showed that areas with vegetation (<28 °C) or close to bodies of water (28 to 32 °C) had the lowest average surface temperatures between the years 1986 to 2016.