Abstract: The effect of digital elevation model (DEM) resolutions on soil-landscape model prediciton accuracy was evaluated by comparing five DEMs (2.5, 5, 10, 20, and 30 m) for a 4.8 km2 watershed in Chongqing. A set of soil-landscape models between soil property and terrain attributes were generated using mutiple stepwise regression. The results showed that the DEM grid resolution had significant effects on the terrain attributes. The mean of specific catchment area and topographic wetness index increased with the decrease of the grid resolution, while the mean of slope and the range of plan curvature and profile curvature decreased with the decrease of the grid resolution. The prediction accuracy of the soil-landscape models was subsequently influnced by the DEM resolutions. The determination coefficient significantly decreased while the mean absolute error and root mean square error remarkably increased when the grid size was lower than 5 m.