Abstract: Micro-topography ecological restoration is one of the important ways to improve the functions of degraded soil, but the driving mechanism of its improvement on soil quality remains unclear. To explore the variation patterns of soil quality and its driving mechanisms under micro-topographic reconstructions, the physicochemical properties of the soil were measured and analyzed by taking the soil at a depth of 0-60 cm in a mining area in western China as the research object. A minimum data set method was applied for comprehensive soil quality assessment, and random forest analysis was used to identify the key driving factors of soil quality. Results indicate that compared with the original slope surface, significant changes in soil physicochemical properties under micro-topography reconstructions were observed, with increases in soil water content (SWC), pH, electrical conductivity (EC), total nitrogen (TN), total phosphorus (TP), soil organic matter (SOM), soil alkali-hydrolyzed nitrogen (SAN), and soil available phosphorus (SAP) contents, yet with a decrease in total kalium (TK) content. Fish-scale pits showed higher content of SWC while horizontal trenches exhibited higher contents of TN, TP, SOM, SAN and SAP. The soil quality index (SQI) of both horizontal ditches and fish scale pits significantly increased. The driving factors of soil quality showed significant variations: fish-scale pits were dominated by TP, SWC and SOM, while horizontal trenches were driven by the synergistic effect of SOM, TP and SAP. This study provides a theoretical basis for the application of micro-topographic reconstructions in ecological restoration and soil improvement in coal mining areas in western China.