Abstract: Hulunbuir Grassland is one of the most important ecological barriers in northern China. However, overgrazing has led to the formation and development of gullies in recent years, leading to degrade and fragment the grassland. There are no studies on the gully erosion in this area. In this study, the field test was conducted on the gully erosion. An area of 136.22 km2 was selected from the Ewenki Autonomous Banner in the center of Hulunbuir Grassland. Eight morphological parameters were measured by laser rangefinders and GPS, including the gully length (L), width (T_W), and depth (D) of each gully. The area (A), volume (V), cross-sectional area (A_c), width-to-depth ratio (R_WD), and length-to-width ratio (R_LW) of the gully were then calculated by ArcGIS software. The gully morphology and distribution were determined to clarify the response of gully erosion intensity to landform change. The volume estimation model was constructed for the gully. The results showed that there was a strong variation in the gully morphology, with a wide range of the eight parameters. The average L, T_W, D, A, V, A_c, R_WD, R_LW were 231.80 m, 3.12 m, 1.26 m, 820.49 m², 1 099.46 m³, 3.39 m², 2.68, 101.93, respectively. The overall shape of the gully was narrow and long, or wide and shallow. The number density and gully density were 1.15 per km² and 0.27 km/km², respectively. Correspondingly, 0.94‰ of grassland was lost, due to the gully erosion. The gully T_W, D, and R_WD were close to the normal distribution. The gullies were distributed mainly in the range of 1.0-4.0 m, 0.5-1.2 m, and 1.0-3.5, respectively, which accounted for 77.9%, 68%, and 77.9% of the total number, respectively. However, the L, A, V, A_c, and R_LW were distributed to the left of the frequency curve in a small range. The gullies were distributed in 0-200 m, 100-600 m², 30-300 m³, 0-3 m², and 3.55-60, which accounted for 63.5%, 56.4%, 46.8%, 57.7%, and 46.1%, respectively. The proportions of hillslope and valley-bottom gully were 55% and 45%, respectively. The development dimension of the valley-bottom gully was higher than that of the hillslope gully, and its L, T_W, D, A, V, A_c, R_WD, R_LW was 1.06-4.78 times higher than those of hillslope gully, respectively. There was a significant effect of topography on the gully development. The gully density (G_D) and gully lacerative degree (G_LD) increased first and then decreased, as the slope increased. The most serious erosion was found when the slope degree was 9°. G_D and G_LD on the sunny and semi-sunny slopes were 4.3 and 3.7 times those of negative slopes and semi-negative slopes, respectively. There were significantly higher correlations of the V with L and A than the rest. Both V-A and V-L relationships were described by the power function. After comparison and justification with the coefficient of determination (R²), relative mean standard error (RMSE), and Nash-coefficient (Ens), the V-A relationship (V=0.059 7A^1.389) was more accurate and effective in predicting the gully volume than V-L relationship. The V-A power function was also suitable for the hillslope and valley-bottom gully, indicating more higher accuracy of estimation. The findings can provide scientific support for the control of grassland degradation caused by gully erosion, particularly in the sustainable development of the regional economy, society, and environment in Hulunbuir grassland.