Abstract: As a large field accumulating salvaged material formed in the opencast working process of coal mining, dump is characterized by unique structure and complex material composition, causing severe soil and water loss. Characteristics of soil erosion on the side slope of dump are significantly different from those on the protogenesis slope. A field investigation was carried out to investigate the soil physical properties and rill morphology of the dump side slope as affected by the vegetation recovery measure used and disposal period served in the Yongli Coal Mine, Inner Mongolia. Bare land and field with vegetation measures of Salix cheilophila., Hippophae rhamnoides Linn., Brassica campestris L. and Astragalus adsurgens Pall with different disposal period (1, 3 and 5 a) were selected. Plot (3 m×12 m) used in the investigation was laid out with an average slope of 35.7°, and the coverage of vegetation varied from 0 to 90%. Soil bulk density was measured using oven-drying method and soil particle composition was determined by a laser diffraction analyzer. The results showed that: 1) The mean contents of clay, silt and sand of the side slope of dump were 7.28%, 22.26% and 70.46%, respectively. The ranges of fractal dimension and soil bulk density were 2.48-2.66 and 1.20-1.44 g/cm3, respectively. As the disposal period increased from 1 to 5 a, the fractal dimension and bulk density all increased no matter for which vegetation measure. The application of vegetation measures effectively reduced the soil bulk density and improved the soil structure of dump side slope. 2) The number of rills, rill density, mean rill depth, mean rill width, rill width-depth ratio and rill split degree fell within the ranges of 5-14, 0.54-2.19 m/m2, 2.50-18.15 m, 5.31-24.81 m, 1.34-2.60 and 0.04%-0.32%, respectively. With the increase of disposal period, the total rill length, rill density and maximal rill length for both the bare land and the plot with Salix cheilophila., Hippophae rhamnoides Linn, Brassica campestris or Astragalus adsurgens Pall increased; whereas, the rill width-depth ratio for the plot with Salix cheilophila., Hippophae rhamnoides Linn, Brassica campestris or Astragalus adsurgens Pall did not increase but reached the maximum when the dump slope had been served for 3 a. 3) The soil erosion modulus increased by 4.11% and 581.28% on bare land and decreased by 27.27% and 50.65% on slope covered by Salix cheilophila, Hippophae rhamnoides Linn. and Astragalus adsurgens Pall while the casting time increased to 3 and 5a. When the casting time increased from 1 to 3a, the soil erosion modulus decreased 43.86% of the slope covered by Salix cheilophila and Salix cheilophila. The soil erosion modulus of slope covered by Salix cheilophila, Hippophae rhamnoides Linn. and Brassica campestris L. and the slope covered by Salix cheilophila. and Astragalus adsurgens Pall was greater than bare land when the casting time was 1 a. The rill breadth-depth ratio of bare land decreased from 2.23 to 1.34 as the casting time increased. 4) The rill erosion amount was extremely significantly correlated with the mean rill depth, maximal rill depth and mean rill width. Among these rill morphological indicators, the total rill length, mean rill depth and maximal rill length exhibited significant correlations with the others and they could be used for evaluation of rill erosion and rill morphology. These findings hold important implications for the eco-recovery of coal mine area.