Abstract: It is difficult to forecast the flood in the near-dam area of Wuqiangxi Reservoir on Yuanshui River system due to the large watershed area, sparse flow stations, and the obvious variation of flood waves propagation after entering into the reservoir area.In order to improve the accuracy of flood forecasting in Wuqiangxi Reservoir area, a XAJ-DCH model was selected to simulate 13 flood events in the near-dam area of Wuqiangxi Reservoir during 2016～2020.Nonlinear Reservoir method and Muskingum method were selected for channel routing.The two different flood forecasting schemes were developed according to the characteristics of the two channel routing methods.The simulation results showed that the deterministic coefficients for the 13 flood events were generally above 0.7,indicating both methods in the XAJ-DCH model perform well and are practical.Compared with the Muskingum method, the Nonlinear Reservoir method can reflect the spatial and temporal variation of flood movement by considering the river section conditions as well as the hydraulic characteristics, and only river roughness needs to be calibrated while other parameters such as riverbed slope, river width and river section length can be estimated based on the Digital Elevation Model(DEM).The Muskingum method needs to calibrate four channel routing parameters, but its simulation result is slightly better than that of Nonlinear Reservoir method.The research results can provide reference for scientific and effective flood forecasting in the reservoir area and improving the forecasting accuracy.