This paper presents a multidisciplinary approach used to assess forest degradation in the Brazilian Amazon based on remote sensing and spatial pattern analysis techniques. The selected study area is located in the Mato Grosso State, which is one of the states of the Brazilian legal Amazon with the highest deforestation rates and with a high concentration of selective logging activities and forest fires. We used an object-based image analysis for mapping degraded forest areas and compared their spatial distribution with that of fragmentation and edge indicators and the distance to roads. Our results show that the majority of these disturbed forest areas occur within a distance of less than 5 km from the main roads, are located between 100m and 5 km from the forest edge, and show higher entropy (used as a measure of fragmentation). However, circa 30% of the degradation occurred in areas considered as “core areas”.