Abstract: Objectives： The drone airborne laser radar technology is widely regarded as a new and effective means of geological disaster investigation in mountainous areas with dense vegetation because it is less affected by the environments, has a unique survey field of view and can “penetrate” vegetation. However, in the existing specifications, there is almost no point cloud density collection standard for geological disasters, and the point cloud density that only meets the mapping requirements is difficult to meet the high -precision digital elevation model production of micro-geomorphic features such as steep ridges and gullies. Methods： To this end, in combination with the characteristics of small scale geological hazards and dense vegetation coverage in Guangxi, this paper studies the DEM quality changes composed of different points cloud density under four large scales, proposes the local terrain complexity of the study area as a quantitative evaluation index, and deduces the optimal ground point density value， Then calculate the recommended value of cloud density by using the laser penetration of different vegetation canopy in the study area. Results and conclusions: The results show that the local terrain complexity can effectively evaluate the micro-geomorphic retention integrity of DEM. Combined with a lot of practical experience, the optimal cloud density value of collection point is obtained by experiment, which established the reference standard of collection point density for airborne LiDAR geological disaster investigation in mountainous areas with dense vegetation in Guangxi.