This study undertook an analysis of the potential of images from the Landsat-8 for classifying the use and soil cover from the far west of Bahia, which has areas of cerrado. For this, 11 scenes of the Landsat-8 system were acquired. To determine the best colour composition, a statistical analysis from the Optimum algorithm Index Factor (OFIS). Supervised classification of Landsat image-8 was carried out through the Support Vector Machine classifier (SVM). The best composition identified was the 7R 6G 5B, which allowed a successful use classification and coverage of the soil of the region. From the supervised, automatic classification by SVM, it was noted that the natural vegetation corresponds to the highest percentage of ground cover in the region, with 66.72%, and along with agriculture, which occupies 27.95% of the area, occupy almost 95% of all area west of Bahia. In addition to the expansion of the agricultural area, the burnt areas occupied 2.78% of the total area, manifesting an alarming result, since it represents a value that is greater than the sum of the areas occupied by classes of urban and water (2.55%). In this way the SVM classifier presented a good result in classification of using and land cover in the region by making necessary just to carry out field visits to perform the verification and validation of the statement of use and land cover produced in this study.