ASTER-satellite data provide useful information for a soil erosion risk model application. An orthorectified FCC and a DEM are the results of a digital photogrammetric restitution. An elevation model can be used to calculate slope angles and the drainage area of each location. So the sensitivity for erosion due to the topographic situation is estimated. In the Mediterranean, rainfall is often of an orographic origin. A linear regression equation taken from literature with the height above sea level as independent variable, and the mean annual rainfall as dependent variable has been adapted for the Chios island. Then the erosivity of the rainfall is estimated. The orthorectified bands are used to make a LULC-classification by means of the maximum likelihood classifier. Before classifying, the image is corrected for differential illumination effects to improve the classification. The different LULC classes have a different sensitivity towards erosion. One non-satellite based information source has been added to the dataset, namely a generalised lithologic map. This map allows tracking of the parent material, which is an important factor determining the soil formation. Different lithologic formations have different erodibility characteristics. When the rainfall erosivity and the sensitivity for erosion due to the topographic situation and the lithology are combined, the potential soil erosion risk is estimated. The potential soil erosion risk is the erosion risk in a virtual world without vegetation. When the sensitivity for erosion of the land cover is taken in account, the actual soil erosion risk is estimated; this is the current soil erosion risk like. At the island of Chios a high potential soil erosion risk is prevalent for most of the land. However the actual soil erosion risk is relatively low for most parts of the island. Vegetation tempers soil erosion. Nevertheless the Mount Epos area is badly degraded and has high actual soil erosion risks.