Abstract: Climate change and urbanization are regarded as the major driving factors for floods in urban watersheds. Many studies have witnessed the impact of climate change and urbanization on streamflow in different regions are different. Identification of driving factors for floods and quantifying the impact of climate change and urbanization on urban flood variation are important significance for floods management in urban watersheds. The Wenyu River basin in Beijing, which has undergone a highly urbanized process, is selected as the study area in this study. The GAMLSS model was developed for different probability of discharges in summer, with seasonal precipitation, air temperature, antecedent wetness, impervious areas and groundwater level as covariates, to study the mechanism of discharge variation in urban watershed. Results showed that all of the different probability of discharges in summer showed the nonstationary characteristics; increasing impervious areas and precipitation are major driving factors for floods in the Wenyu River basin. Different probability of discharge has different driving mechanism, smaller discharges with probability higher than 70% was mainly affected by changes of land use on ground surface, while the greater one, especially the extreme floods below 45% probability were mainly influenced by precipitation.