To understand how the produced flood boundaries from hydraulic modeling can be affected by the resolution of the Digital Elevation Model (DEM) and the roughness values used, and how different performance measures can affect the choice of most optimal models, different simulations were conducted using varying pairs of DEM and Manning’s n. Each result was then compared to the reference flood data and analyzed using different performance measures that account for the sizes of inundation (F-statistics), and the distance between the modeled and reference flood (disparity measures). The results showed that the different pairs of DEM resolution and Manning’s n produced different performance scores and flood extents. High-resolution DEMs performed generally better with higher resolution DEMs, while lower resolution DEMs received better performance when paired with lower Manning’s n. The different model performances also produced different combinations having the most optimal performances depending on the assumptions they applied in the computations. Additionally, the results showed that models from coarser resolution DEMs can produce higher maximum scores than higher resolution DEMs. However, when the maps, cross-sectional profiles, and the water surface elevations from the coarser resolution DEMs were compared with the higher resolution DEMs, they are more inaccurate in representing the terrain, flood boundaries, and the water depths. Thus, interpreting model results only based on performance measure, particularly for lower resolution DEMs, may give incorrect conclusion on the optimality of the result. It is important that the flood extent maps, cross-sectional profiles, and the derived water depths are analyzed together with the performance scores to gain better understanding of the reliability of flood model results.