Soil erosion is a very critical form of land degradation resulting in the loss of soil nutrients and downstream sedimentation of water storages in the highlands of Ethiopia. As it is technically and financially impossible to conserve all landscapes affected by erosion, identification of priority areas of intervention is necessary. Spatially distributed erosion models can help map landscape susceptibility to erosion and identify high erosion risk areas. Integration of erosion models with geographic information systems (GIS) enables assessing evaluate the spatial variability of soil erosion and plan implementing conservation measures at landscape levels. In this study, the Revised Universal Soil Loss Equation adjusted for sediment delivery ratio was used in a GIS system to assess landscape sensitivity to erosion and identify hotspots. The approach was applied in three catchments with size being 10–20 km2 and results were compared against quantitative and semi-quantitative data. The model estimated mean soil loss rates of about 45 t ha−1 y−1 with an average variability of 30% between catchments. The estimated soil loss rate is above the tolerable limit of 10 t ha−1 y−1. The model predicted high soil loss rates at steep slopes and shoulder positions as well as along gullies. The results of the study demonstrate that knowledge of spatial patterns of high soil loss risk areas can help deploy site-specific conservation measures