Based on an ETM, which utilized remote sensing software and geographical information systems software to construct a number of channel compositions, and by employing geometry correcting technology with the Universal Soil Loss Equation (USLE) as a pair of theory foundations, we used map algebra methods to assess the sensitivity of soil erosion in ecologically functional areas in Dongsheng, Erdos. Within this region were typical grassland hills areas, typical grassland districts, typical grassland desertification areas, desert grassland districts and urban residential districts. We classified the influence of precipitation, soil, topography and vegetation upon factors of soil erosion into five different degrees. As a result, we concluded research with an assessment map of each factor. We also drew integrated assessment maps of the sensitivity of soil erosion by overlapping these functions with ArcGIS. Through these assessments, clear possibilities for light, area coverage and spatial distribution patterns in soil erosion sensitivity came to light. Steppe hilly area of soil erosion sensitivity index SSj was 4.2, typical steppe plain area was 4.0, typical grassland desertification area was 4.6, desert grassland area of soil erosion sensitivity index was 5.0; This report also assesses the effective countermeasures controlling said soil erosion. The result will provide an important support base for partition management of ecological environments, the sustainable use of land resources, effective control of soil erosion, which will be of wide interest in the future.