Abstract: Aiming at the ecological damage problem in the historical legacy mining areas of Gongji Community in Chenzhou, Hunan, a multi-source data modeling system integrating laser point cloud, remote sensing image, and ground measurement is constructed to complete the spatial identification and classification extraction of terrain, hydrology, soil, and landscape elements. Based on the DEM model, restoration units are divided, construction paths and operation sequences are optimized, slope support, soil improvement, and vegetation reconstruction schemes are matched. HRB400 grade threaded steel anchor rod and ZYG-450 type hydraulic spraying equipment are adopted for collaborative operation. Through comparison, areas with excessive slope are decreased by 71.6%, the soil pH is increased to 6.5, and the organic matter content is increased to 2.45%, which is more than 163% as much as before restoration, and the ecological plaque connectivity index reaches 0.61. The results verify the effectiveness of multi-source data fusion in terms of improving restoration accuracy and efficiency.