The appraisal of farms market value is widely used by the Brazilian Federal Government as a tool for obtaining land. For this, it is necessary to define its characteristics. The use of vegetation indices for determining the ground cover is an appropriate tool for this purpose. This study aims to evaluate the potential of SFDVI (Spectral Feature Depth Vegetation Index) in land cover discretization and compare their performance with two other indexes traditionally used for this purpose. Therefore, we use a RapidEye image and selected the NDVI and EVI indexes to compare their individual ability to identify soil, grasslands, gallery forest, savannah and water. The result of applying the indexes showed that SFDVI presented a higher potential, especially for targets with dense vegetation cover. In addition, the similarity was tested using the Mann-Whitney test and verified that despite the three presenting high correlation, they are not similar.