Abstract: To address the problems associated with increases in decision variable dimensions due to the need to consider gate operations and water level control at the end of scheduling in the joint optimization scheduling of hydropower station flood discharge and power generation, this paper proposes a hierarchical optimization strategy for units and gates. All gates are aggregated into a virtual gate. The upper-level model conducts flood discharge and power generation optimization scheduling for units and the virtual gate, whereas the lower-level model distributes the total flow of the virtual gate among the actual gates to achieve problem dimensionality reduction. Moreover, a flow-classified water level control strategy is proposed: joint optimization scheduling is carried out when the inflow is high, and water is discharged according to the water level control rules when the inflow is low to ensure a reasonable drawdown of the water level at the end of scheduling. Taking the Xiajiang Reservoir as an example for analysis, the results show that compared with the traditional joint optimization model, the hierarchical model reduces the peak flow from 17,490 m³/s to 16,648 m3/s, increases the peak cutting rate from 22.3% to 26.0%, and increases the power generation from 5.58 million kWh to 5.67 million kWh. After coupling with the water level control strategy, the peak cutting rate increases to 26.6%, and the final water level at the end of scheduling decreases from 46.00 m to 45.47 m. In conclusion, the results provide a technical reference for the joint optimization scheduling of hydropower station flood discharge and power generation.