Abstract: The goaf in metal mines is one of the main hazards threatening the safety of mines because of many factors and complex shapes. In order to evaluate the present situation of the goaf rationally and put forward effective control measures， the first task is to accurately describe the three-dimensional shape elements of the goaf. Accurate detection technology of goaf based on 3D laser scanning has been mature， which provides data basis for fine 3D modeling of goaf. Traditional forward modeling technology is cumbersome， inefficient and requires high data quality and personnel technical level. In order to improve the efficiency of goaf 3D modeling and reduce the modeling difficulty， the key points and difficulties of laser point cloud data processing and model reconstruction technology in metal mine goaf are summarized and discussed based on the principle of reverse modeling technology， and a complete workflow for 3D modeling of complex goaf based on reverse engineering technology is proposed. Finally， the proposed method is applied to the analysis of typical goaf problems in metal mines and the design of reinforcement， blasting and other treatment schemes.The study results show that reverse modeling method obtains the 3D entity model based on the surface that reconstructed directly from the point cloud information， and avoids a lot of repetitive line-surface and surface-body operations in forward modeling. Besides that， this high-efficiency and high-precision modeling method could provide technical support for the investigation， assessment， disaster mechanism analysis and treatment scheme optimization of goaf in metal mines，has broad application prospects.

Key words: Metal mines, Goaf, Accurate detection, Reverse engineering, Modeling efficiency