The system response error correction method was introduced and improved, and a Muskingum continuous routing error correction method based on system response was proposed to solve problem of the declining accuracy of the traditional Muskingum method in river basin flood forecasting by using the Xin’anjiang model, due to the continuous construction of river channel water conservancy projects such as runoff hydropower stations and ship locks in recent years. A correction coefficient was calculated based on the difference between observed and calculated discharges at the forecast section, and was applied to the nodes of various upstream river sections to conduct real-time dynamic corrections of the initial values of river confluence. Taking the main river channel of the Fuchunjiang Basin, which is affected by cascade ship locks, as a case study, the results show that the flood forecasting accuracy after correction is significantly improved. The correction method has the best correction effect on flood peaks, and it also has a remarkable correction effect on double peak or multipeak floods, especially on those flood processes greatly affected by water conservancy projects. The traditional Muskingum method, combined with the system response error correction method, can effectively reduce the impact of river channel water conservancy projects on flood forecasting, improve the accuracy of river basin flood forecasting, and solve the problem of asymmetric information dimensions between information sources and variables to be corrected in the traditional system response error correction method.