Abstract: During the past decades, frequent flooding disasters in urban areas resulted in catastrophic impacts such as human life casualties and property damages especially in piedmont cities for their specific topography. This study develops a piedmont urban flooding model for the Huangtaiqiao drainage area based on SWMM. The sub-catchments in this study area are divided into mountainous area, plain area and main urban area according to the variations in surface topography. The impact of different routing modes and channel roughness on the simulations is analyzed under different types of scenarios, and a genetic algorithm is used to optimize model parameters. Results show that the simulation is poor (with a mean Nash coefficient of 0.61) when using the traditional routing mode in SWMM model, which usually ignores the difference in surface topography in piedmont areas. However, when the differences in routing modes, percent routed and channel roughness are considered, prediction accuracy is significantly increased (with a mean Nash coefficient of 0.86), indicating the significant affect of the difference in surface topography on the simulations of piedmont cities.