When the grassland ecosystem is under investigation, soil quality indexes (SQIs) are constructed to evaluate soil quality substantially, especially in desert grasslands and other ecologically fragile areas. This research used the total dataset (TDS), minimum dataset (MDS), approaches dealing with selecting better indexes, and scoring methods utilizing linear and nonlinear expressions to assess the typical desert grasslands in Yanchi County, Ningxia, China. The utilization of four different lands such as forestland (FL), shrubland (SL), natural restoration grassland (GL), and abandoned farmland (AL) in the study area were researched. Physical, chemical, and biological indicators, a total number of twenty, were measured. Principal component analysis and norm values were used to select indicators based on the MDS. The results suggested that nine soil indicators, namely, soil water content (WC), total soil porosity (TP), percentage of soil sand (sand), percentage of soil clay (clay); soil organic carbon (SOC), total nitrogen (TN), available nitrogen (AN), urease activity (UA), and catalase activity (CA) were selected for the MDS. The distribution of the SQI in the types of land use was similar concerning the two evaluation methods. The nonlinear scoring method utilizing the MDS was found the most proper to compute the SQI since the maximum F statistics, coefficient of variation (CV), and correlation results are obtained. The SQI outcomes that were ranked concerning the types of land use were found to be shrubland (SL) > natural restoration grassland (GL) > abandoned farmland (AL) > forestland (FL). In the four types of soil, shrub afforestation can be used as a beneficial ecological measure to restore the soil quality of typical desert grasslands in the research area.