Remote sensing technology and machine learning algorithms are used for quick and accurate land use monitoring in mining areas, enabling ecological environment monitoring and scientific evaluation. The study focuses on Dawukougou in Helan Mountain. The GF-2 remote sensing image is used with object-oriented classification and machine learning to classify land use in two high-resolution remote sensing images before and after ecological restoration of abandoned mines. Statistical analysis is done on the change in ground object area in the research area. The results show that the area of vegetation, bare land and mining area has changed greatly after ecological restoration in the study area. The area of the mining area changed from 34.64 km2 to 8.7 km2, a decrease of 25.94 km2. The bare land area changed from 231.12 km2 to 255.71 km2, an increase of 24.19 km2; the vegetation area increased from 5.19 km2 to 6.49 km2, an increase of 1.84 km2. After ecological restoration, there is a clear spatial correspondence between the bare land and the area with increased vegetation and the area with reduced mining area. The reason why the vegetation area increased slightly and the bare land area increased significantly after ecological restoration in the study area is that the local natural geographical conditions are harsh and the ecological restoration project is completed soon. The research results can provide reference for the ecological environment monitoring and ecological restoration effect evaluation of abandoned mines, and provide technical support for the ecological stability and social and economic sustainable development of mining areas.