•An optimization model is proposed by coupling the SCS-CN, NSGA2, and MLP algorithms.•This model improves the spatial balance of retention water volume.•The risk of urban waterlogging is mitigated through the runoff-sharing effect.•This model is better than other optimization models for impervious surfaces. The expansion of impervious surfaces during urbanization, as well as the irrationality of their spatial layout, is one of the main causes of urban waterlogging. Optimizing the spatial layout of impervious surfaces through urban renewal is important for urban waterlogging prevention. The differences in hydrological conditions among different hydrological units affect the optimization efficiency of the spatial layout of impervious surfaces in different hydrological units. Therefore, this study proposes an effective optimization model for impervious surface spatial layout, employing a hybrid approach that couples the Soil Conservation Service Curve Number (SCS-CN) model, nondominated sorting genetic algorithm 2 (NSGA2), and multiple linear programming (MLP) algorithm. This approach is applied in a case study of Guangzhou, China. The results show that (1) compared to the MLP-SCS model, the enhanced model not only provides reduced runoff coefficients but also achieves a better spatial balance of the retention water volume. (2) Compared to the conditions before optimization, medium- and high-density impervious surfaces significantly change after optimization, and the connectivity between patches is reduced. The optimized layout alters the original gradient connection method, preventing the aggregation of runoff from high-density impervious surfaces and thereby reducing the risk of urban waterlogging during heavy rain. This study can provide a reference for optimizing the spatial layout of impervious surfaces aimed at urban waterlogging prevention in urban renewal planning.