Abstract: Coal mine industry has always been an important part of China's energy industry. Under the general trend of developing digital mines, it is particularly important to break through the traditional two-dimensional drawing management and realize the development of monitoring and management system to information technology. In this process, one of the basic tasks is the establishment of coal mine industrial square model. Traditional modeling methods of coal mine industrial squares are often limited by data acquisition methods and accuracy, which cannot fully and accurately reflect the actual scene. This study proposed a method of integrating multi-source data, aiming to make full use of the advantages of UAV oblique image and airborne Lidar point cloud, and make up for the defects of single data, so as to improve the accuracy and comprehensiveness of modeling. The results show that: Uav tilt photogrammetry technology has dead angles in aerial photography, resulting in partial missing of buildings, interrupted connection, glass reflection and other phenomena causing local model texture glitter and distortion, while airborne lidar technology has the problem of incomplete data at the bottom of buildings. The data model after fusion can solve the problems of the above two technologies and improve the digitization level of coal mine industrial square.