The study attempts to evaluate the impact of DEM source (AW3D30 DEM, CartoDEM v2 R1, SRTM v4.1 DEM and ASTER GDEM v2), DEM resolution (30 m to 1000 m), resampling approaches (nearest neighbor, bilinear interpolation, cubic convolution, majority) and area threshold (1500 Ha, 10,000 Ha, 25,000 Ha, 35,000 Ha, 50,000 Ha) on hydrological model (SWAT) simulated outputs. A methodological framework by two criteria: (1) DEM quality assessment and (2) river network delineation capability of DEM were developed for identifying best DEM among the considered DEMs for baseline scenario. It is found from the study that AW3D30 DEM best represented the terrain of the catchment among the evaluated topographic models with a least RMSE value of 7.44 m. Further AW3D30 DEM had the best river network extraction capability with a minimum RMSE value of 44.52 m in comparison with reference network. All the DEM scenarios were found to be insensitive for surface runoff. Ground water flow, evapotranspiration, potential evapotranspiration and water yield estimates did not show any sensitivity to DEM scenarios but soil water content showed its sensitivity to area threshold scenario. In water quality estimates, all DEM scenarios were found to be highly sensitive to sediment yields in comparison to total nitrogen and total phosphorus.