Due to the flat topography, complex water system, and strong human activities, the hydrological process is complex. With global warming in recent years, extreme precipitation events have increased and floods are frequent, it is necessary to investigate the hydrological simulation and its response to extreme precipitation in the plains. By improving the traditional river and sub-basin extraction through reasonable generalization and spatial discretization of the river network, the SWAT model was applied to simulate runoff process in the Lixia River Plain during 2013-2020, and the runoff changes under different extreme precipitation scenarios were analyzed. The results show that the SWAT model has a good applicability in the simulation of monthly runoff during the flood season in the Lixia River Plain, as the correlation coefficient R2 and Nash efficiency coefficient NSE were above 0.75 and the relative error |Re| was less than 5% at Funing (She) station. The daily average runoff varied widely and the peak changed obviously under different extreme precipitation scenarios, with the change rate of 3.4%, 10.2%, and 14.2% under the once in 20-year, 50-year, and 100-year extreme precipitation events, respectively. While the monthly average runoff changes were small, and impact time was short. The study could provide references for water resources management and response to floods in the Lixia River Plain.