China is one of the countries that seriously suffered desertification in the world, and it is very meaningful to develop a quantitative method to assess desertification at large scale. In this study, the MODIS images were selected as the data resources, and NDVI, land surface albedo, soil water index (the reflectance of MODIS band 7) were selected as the indicators for assessing desertification. Based on building the indicator rule sets of different desertification grades in different sub-regions, the authors developed a quantitative method for desertification assessment by using decision tree model. The results showed that, the method developed in this study that can reflect the heterogeneity of land surface at large scale, and the overall accuracy of the method can reach 85.5%, which was suitable to assess desertification at large scale. Based on using this method to assess the desertification in farming-pastoral region of north China in 2000 and 2010, the authors found that the areas of lands that experienced desertification reversion and desertification expansion were almost consistent, and the spatial distribution of these regions existed obvious differences.