The aboveground biomass of vegetation in desert areas serves as a crucial indicator for monitoring land desertification and extracting desert vegetation information using remote sensing techniques. In this study, the Minqin County of Gansu Province was selected as the experimental area and Sentinel-2 images were used as the data source. We constructed estimation models (unitary linear, exponential, logarithmic, and binomial models) for the planted index and the aboveground biomass of vegetation, which were measured by us. These models include five vegetation indices: ratio vegetation index (RVI), normalized difference vegetation index (NDVI), difference vegetation index (DVI), soil-adjusted vegetation index (SAVI), and optimized soil-adjusted vegetation index (OSAVI). The aboveground biomass in the study area was estimated using the selected optimal model. The results demonstrated that SAVI had the highest correlation with the aboveground biomass (r = 0.79) compared with RVI, NDVI, DVI, and OSAVI. The binomial model based on SAVI was the best model (R2 = 0.76) for the aboveground biomass estimation in the study area, with higher accuracy (R2 = 0.73, RMSE = 0.12). In the Minqin County, the relatively dense areas of vegetation were mainly distributed in the four major irrigation districts (Hongyashan, Huanhe, Changning, and Nanhu), the surrounding area of Qingtu Lake, and the northwest region of Hongshagang Town, whereas the vegetation in other regions was relatively sparse. The proportions of nonvegetation area <[0.005 kg·(100m2)-1], low vegetation area [0.005-0.2 kg·(100m2)-1], medium vegetation area [0.2-0.5 kg·(100m2)-1], and high vegetation area [>0.5 kg·(100m2)-1] were 66%, 21%, 5%, and 8%, respectively.