In the field of auxiliary airborne navigation, the current Lidar-based point-cloud terrain segmentation method has low real-time performance and poor adaptability to different terrains.To solve the problem, this paper proposes an adaptive point-cloud terrain segmentation method, which combines the CSF algorithm with the fitting of moving surfaces.This method calculates the hardness value of the cloth through the gridding and recognition of terrains to adaptively segment the point clouds, and then uses the cloth model established in the segmentation process to generate terrain grids for helmet-mounted display.Experimental results show that, compared with the classic terrain segmentation methods released by ISPRS, this method has the advantages of strong adaptability to different terrains and high real-time performance on the basis of ensuring certain accuracy, which has certain value in the application of the Lidar to auxiliary helicopter navigation.