Natural resources in mountainous terrain are profoundly affected by land degradation due to anthropogenic pressure. Being a mountainous landscape, soil erosion by water is considered a critical problem in the Himalayan terrain. Soil erosion is being accelerated in this region due to intensive deforestation, terrace farming on steep slopes, large-scale road construction, mining, etc. Mapping and assessment of erosion prone areas enhances soil conservation and watershed management. Field size, land use, vegetation, soil, slope, and geomorphology contribute to the complexity and heterogeneity of natural landscapes. To evaluate soil loss for each landscape level (field- plot, watershed, and landscape scale), a hierarchical approach of spatial risk modeling was used in this study. To estimate the relevant soil loss, spatial soil erosion risk modeling exercise that was based on the Morgan, Morgan and Finney Model (MMF), with the aid of remote sensing and GIS, was carried out for a small subwatershed of the larger Sitla-Rao sub-watershed in the Doon Valley, Uttaranchal State, India, where about 60% of the total area is mountainous. Satellite image #IRS-P6 LISS IV (5.8 m resolution) was visually interpreted for land use, land cover, and physiographic soil mapping. A slope map was generated by using DEM, derived from terrain topographical data (Survey of India). Rainfall data are acquired from the local weather station. Intensive field work was carried out in the watershed using global positioning system (GPS) for collecting input data such as crop cover management factor and soil hydrological characteristics. Thirty-three different soil profiles representing various physiographic soil units along a selected transect in the watershed were studied for morphological characterization and taxonomic classification. Soil orders are found to be Entisols and Inceptisols. Collected samples were analyzed to obtain input data for MMF model. Spatial modeling was carried out in a GIS environment. Soil erosion predicted by the model was found to be comparatively higher for areas with open scrub vegetation growing on steep to very steep slopes in the hilly areas. Cropland areas contributed to moderate, moderately severe, and severe erosion. The lowest soil loss was predicted to be for areas in forest. Various approaches were used to validate the predicted results at each level of landscape. For field-plot scales, field run-off plots were monitored during the monsoon season for soil erosion. For watershed and landscape scales, field surveys were conducted for two purposes: 1) assessment of erosion hazards by observing and measuring the occurrence of erosion features, and 2) Assessment of Current Erosion Damage (ACED).Incidences of erosion features were recorded at each scale and their mean values were analyzed statistically. “Erosion-topo-sequences” covering entire cross-sections of a slope, ideally from hilltop to valley floor were created by using ACED. Comparison of the erosion predicted by the model with the erosion observed in runoff plot was made. Results obtained from plots highlighted that runoff varies with both crop cover-management factor and slope factor. Forests have a positive impact on controlling excess runoff. The modeled soil losses were also compared with the soil erosion data obtained from field surveys at the watershed and landscape scales and it was found that erosion varies 20-35%. Soil loss in the experimental watershed was determined to be moderate, since soil loss up to 25 tonnes ha^-1 yr^-1 is considered tolerable in mountainous area (Morgan, 1986) where the natural soil loss rate is high.