In order to explore the impact of baseflow separation on the uncertainty analysis of urban storm water simulation, a storm water management model(SWMM) is developed taking downtown area of Jinan City as the study area. Observation data of 7 floods occurred during 2005 and 2010 before and after baseflow separation is used to calibrate and validate the model. The generalized likelihood uncertainty estimation(GLUE) method with root mean squared error used as its basis is applied in uncertainty analysis for the simulation results. This paper also discusses the impact of whether data is collected before or after the baseflow separation on the uncertainty of simulation results. The results show that the average baseflow of rainstorms to peak discharge is 13.38%, which has a significant impact on flood process simulations. The GLUE method can effectively extract the parameter set with high precision. Discharge after the baseflow separation is more appealing as the basis for the verification of model compared with that before the baseflow separation, where 90% confidence interval of flood process is 100% higher, average offset amplitude is 42.60% lower and symmetry is 87.19% lower. The baseflow separation can effectively reduce the uncertainty of the simulation result, improve the performance of flood forecasting and help obtain flow forecast interval with more accuracy and symmetry.