Abstract: Middle Route of the South-to-North Water Transfer Project was constructed into a large-scale series canal system in a centralized automatic control mode for an operation management. There are the striking characteristics of long water transmission lines, and large scale of water transfer. Specifically, the height difference is 100 m from Danjiangkou Reservoir to Beijing. The water distribution can be achieved by adjusting the control gate using artesian water delivery, without any online regulation reservoir along the canal line. Therefore, a highly accurate adjustment is necessary for the flow regulation and control of the gate group during the operation and scheduling process, in order to realize the safe timely water delivery in an appropriate way. Many difficulties have arisen on the hydraulic control and dispatch of the main canal, due to numerous buildings along the main canal, while, the variations in water demand of each water diversion gate. Furthermore, a large hysteresis of hydraulic response usually occurred, due to the limitation from the propagation speed of the water wave. Accordingly, the change in the flow of any water diversion or control gate along the canal line can cause water level fluctuations within a certain channel range, showing a strong coupling effect. As such, the risk of ice jam can increase significantly, because of large hydraulic fluctuations during the freezing period, particularly on large-scale series canal systems. How to reduce hydraulic fluctuations through joint dispatching of gate groups can efficiently restrain the occurrence of ice jams during the freezing period in this case. In this study, taking the minimum deviation of water level as the objective function, an adjustment system was designed, including two control links, a conventional PI controller, and an optimization controller, while, combining with a genetic algorithm, a simulation model for the optimal dispatching of the sluice gate group was established suitable for the frozen period of the canal, considering the flow restriction between the canal pools, gate opening, and adjusting rate constraint. An optimization controller was comprehensively demonstrated, according to the verified modeling effect and the parameter sensitivity, based on the simulation experiment of control gates in the canal system between the ancient and the north Juma River currently. The simulation results show that the maximum deviation of water level can be significantly reduced under the simulated operating conditions, when adding an optimization controller in the system, compared with the only PI controller. Specifically, the maximum deviation of downstream water level decreased by nearly 36%, whereas, the recovery time of system was nearly 2.9 h ahead of time, indicating that the proposed model has a positive effect on suppressing the excessive hydraulic response, and stabilizing the water level during the freezing period. This will be beneficial to reduce the risk of ice jam. The reason was that the decrease in the deviation of the water level can be implemented via the flow control at the inlet and outlet of each channel pool, while better coordinating the need for storage adjustment between channels. The maximum opening of each gate can be reduced, but the adjustment amplitude of a single gate opening increased for some channels. The adjustment of gate opening degree can be feasible, due to the constraint of gate opening and adjustment rate. In the solving process of genetic algorithm, the specific value of random disturbance flow was recommended relatively small, due possibly to a negative effect on the optimization direction. In the constraint parameter of flow range, d, there was only a limited influence on the fluctuation range of water level, and the number of optimization iterations.