The failure of the Fundão Dam, in Mariana (MG), on November 5, 2015, is considered “the greatest environmental tragedy in the country's history”. In this way, this research has the objective of analyzing the impacts of the failure of the Fundão dam in the municipality of Mariana-MG on the vegetation cover, by means of remote sensing techniques and analysis of digital processing of orbital optical images. To analyze soil cover, the Unmixing Espectral Linear Model (UELM) was used to separate soil, shade and vegetation classes. Subsequently, the Artificial Neural Network (ANN) classification method was applied. The results showed a loss of 13.02% of the vegetation, about 1,289 ha and a reduction of 68.57% of shade (water), approximately 1,347 ha. The UELM showed to be effective in the separation of each image fraction, being an important stage for the success of the classification.