Abstract: A gully is a kind of earth erosion resulting from many external forces. Given the huge labor intensity required in traditional gully monitoring methods and the great influence of complex terrain, a terrestrial laser scanning (TLS) method for gully monitoring and erosion estimation in the area of sparse vegetation is presented in this paper. A typical gully on the east bank of Guanting reservoir in Hebei Province was taken as a case study for TLS-based gully monitoring and erosions estimation using three phases of TLS data collected in a two year period. Applying TLS point cloud acquisition, registration, filtering, resampling and surface fitting, the surface models of the gully were reconstructed at different spatial resolutions, and the terrain information of the gully was extracted from the surface models. Erosion was estimated using the Yang Chizhong Filter at different spatial resolutions with resampled TLS point clouds, and compared. The results show that: 1) the estimated erosion was more stable and accurate when the spatial resolution of resampled TLS point cloud was equivalent to the size of concavo-convex structure (2~6 cm) of the gully, 2) the general elevation of the gully decreased 2~20 cm during the monitoring period, 3) the gully surface with the largest curvature showed the most significant erosion, and.