Abstract: China's densely vegetated mountainous regions are prone to frequent geo-hazards, and existing DEM visualization techniques are inadequate for depicting the micro-topography associated with these hazards. This paper proposes an enhanced display algorithm for micro-topography based on DEM terrain profile features. Inspired by the random unstable spectral signal processing, this algorithm proposes to treat terrain profiles as signals. The fundamental approach involves segmenting the DEM data obtained from LiDAR into terrain profiles, fitting terrain curves to each profile, and processing the differences between the fitted and original terrain curves. The micro-topographic features obtained from these differences are then amplified by a specific coefficient and used to reconstruct the DEM profiles. This process is applied to all DEM profiles to obtain the enhanced DEM data. The study selected one landslide area as research sites and applied the proposed algorithm to enhance the landslide DEMs. The results show that the enhancement algorithm proposed in this paper is less affected by the direction of sun light, which can effectively improve the visibility of micro-topographic features on the landslide DEM, and is more conducive to the visual interpretation of geo-hazards. The results demonstrate the algorithm's validity and provides effective technical support for the early identification and interpretation of micro-topographic features in geo-hazards.