Abstract: To address common issues in urban catchment-scale rainfall-runoff simulations, such as inadequate topographic and drainage data, concentrated flooding on the road area, and the poor calculation accuracy or low calculation efficiency of the existing flooding models, a partition adaptive model based on the concept of plot generalization simulation and road networks fine simulation is proposed. For plots with insufficient data, sub-catchment units are divided, and the hydrological generalization method is used to calculate the runoff generation and concentration process. For the road area, fine grid division is carried out, and the two-dimensional hydrodynamic method is used to simulate the flow movement process. Both methods are coupled with the pipe network model to jointly drive the pipe network model to perform the confluence process calculation. The core demonstration area of Fengxi New Town in Xixian New Area is used as the research object to build a partition model. Besides, the full-distributed model with the study area discretized by fine grids and the semi-distributed model with the study area discretized by coarser sub-catchments are applied in the same area for comparison. The results show that : ① The partition model has the advantage of comprehensively considering the drainage of plot drainage and runoff process, and can simulate the inundation process in low-lying areas of road under the combined effect of surface runoff and pipe network overflow, which conforms to the actual rainfall flow motion. ② Compared with the semi-distributed model, the partition model can better simulate the process of water accumulation in the road area by using fine grids. ③ Since the calculation task of the partition model is less than that of the full distribution model, simulation time can be reduced by 28.2%—73.5% under different return periods. The new partition model effectively overcomes limitations of data conditions and computing power, and can provide a new coupling simulation idea for urban flood simulation.