Abstract: For the multi-parameter calibration problem of the complex cascaded open channel water transfer project, the traditional trialand-error method needs to obtain the optimal combination of parameter values based on the hydrodynamic model, which requires a large amount of computation. Therefore, this paper proposes a multi-parameter calibration method based on a genetic algorithm, which transforms the multiparameter calibration problem in the hydrodynamic model into an optimization problem with the calibrating parameters as the state variables. Then the combination calibration of the parameters is obtained by a genetic algorithm. The method is used to determine the roughness of the three channels and the flow coefficients of the three undershot gates in the “Xiheishan gate-Fenzhuanghe gate” section in the Middle Route Project of the South-to-North Water Transfer Project. The results show that the optimal combination calibration of the parameters can be obtained by one optimization process, the water level obtained by the simulation model by using the calibrated parameter values are relatively close to the measured water levels, with the average deviation less than 2 cm.