Climate change alters hydrologic cycle and reservoir inflow and brings about conflicts between different objectives in reservoir operation. A significant issue in reservoir management is how to adjust reservoir operation rules accordingly through research considering the possible future changes in runoff and hydrologic cycle. In this study, we have developed a multi-objective optimization model for operation of the Danjiangkou Reservoir considering possible future runoff changes. To predict future runoff in this model, the results of the global climate model (GCM) are downscaled by a statistical downscaling model and then inputted into a calibrated variable infiltration capacity (VIC) model. And multi-objective operation rules can be derived using a non-dominated sorting genetic algorithm II (NSGA-II) method,combining the design operating rule curves with the predicted future inflow series and considering the tradeoff between power generation and water supply. Results show that this reservoir needs to modify its existing operation rules to adapt to the future runoff changes and the multi-objective operating rules so obtained are effective in mitigating the conflict between water supply and power generation and improving comprehensive water utilization. In general, power generation-oriented and water supplyoriented rules will produce different results in the objectives, among which the water supply-oriented rule is able to increase water supply capacity significantly (by 18% in the Danjiangkou case).