Abstarct: In recent years, climate change and urbanization have reduced the permeability and time of concentration in urban areas. Therefore, it is obvious that if there is no suitable program for management of precipitation, the resulting runoff can cause severe floods in these areas. In addition to traditional approaches on the surface runoff management, e.g. collecting and draining urban runoff and transferring it safe downstream points, it is also possible to use low Impact Development (LID) methods as a novel approach to quantitative and qualitative urban runoff management. The purpose of various LIDs is to reduce the impact of the development, maintain or restore the natural hydrological conditions of the catchments to the pre-urban development. By using LIDs, it is possible to provide suitable conditions for reducing the contaminant in urban runoffs and increasing the possibility of feeding groundwater aquifers and reused runoff as a source for providing non-drinking water demands. In this research, after modeling the drainage network of the case study in the Storm Water Managemanet Model (SWMM), the flood points and flood volume produced for each of the 45 probable precipitation scenarios were determined. Then potential retention ponds as a LID approach were identified using Google Earth and were confirmed by field surveys. As optimal design of these retention ponds is one of the main engineering concerns, the combination of rainfall-runoff simulation model with an optimization model, i.e. NSGA-II algorithm was used to determine the optimal depth of proposed retention ponds. The outputs of the simulation-optimization model resulted in the Pareto optimal solutions and with regard to economical aspects or the amount of flood reduction, an optimum solution can be selected. Also, by calculating the Investment Priority Index (IPI) for each of the retention ponds available in Pareto optimal solutions, it is possible to decide on the priority of their construction. One of the uncertainties in the rainfall-runoff modeling is the estimation of rainfall losses, e.g. infiltration, which can be estimated using the SCS-CN method. Using fuzzy logic and introducing CN as a fuzzy number can deal with the existing uncertainties. In this study, to assess the effect of CN changes on flood rate and optimal depth of retention ponds, the corresponding CN values of α-cut 0, 0.5 and 1 were considered in the simulation-optimization model. The results showed that with a 42% increase in the CN coefficient from dry to wet conditions, the number of flood points increases by 88%. Also, the frequency of flooding in flood points and the total amount of flood volume increase by 3 and 10.86 times, respectively. Accordingly, in addition to the hydraulic conditions, hydrological parameters and their uncertainties, such as the rainfall intensity and temporal distribution and infiltration, should also be considered in design of urban runoff systems. It can also be deduced that retention ponds have a positive effect on flood management, e.g. in our case study the total flood reduced by an average of 80.2 to 85.6 percent.