The complex and diverse nature of coal mining sites, including different landforms and working conditions, presents challenges for rehabilitation efforts. To address this, we conducted a comprehensive experimental study focusing on microbially induced calcium carbonate precipitation (MICP) remediation, considering the fracture characteristics of coal mining sites. The MICP-restored samples were subjected to confined/unconfined compressive strength, uniaxial/triaxial permeability, and souring tests to assess their restoration efficacy. The results showed that under similar mining conditions, the average depth of parallel fractures was 0.185 m for loess ridges, 0.16 m for the valley, and 0.146 m for the blown-sand region, while the average depth for boundary fractures was 0.411 m for loess ridges, 0.178 m for the valley, and 0.268 m for the blown-sand region. Notably, parallel fractures showed negligible filling in all landforms, whereas boundary fractures in the blown-sand region were completely filled with wind-deposited sand. The valley landform was filled with alluvium and wind-deposited sand, whereas the loess landform was filled with wind-deposited sand and loess. MICP-restored soil samples in all landforms achieved a strength comparable to remolded fracture-free soil samples. Across all landforms, the maximum permeability coefficient of MICP-restored soil samples closely matched that of remolded fracture-free soil samples. Under similar topographic and rainfall conditions MICP restorations scoured 31.3 g on blown-sand region, 19.3 g on loess ridges, and 17.6 g on valleys. These research findings provide an experimental foundation for MICP repair of coal mining ground fractures.