Since there is no consensus about how to assign weights in Global Circulation Model (GCM) ensemble, this paper uses the Dempster-Shafer (DS) evidence theory to synthesize the following three the basic probability assignment (BPA) methods: the equal probability,the probability for statistic characteristics of the mean annual inflow,and the probability based on the relative monthly inflow variation. Then,DS evidence theory-based Adaptive Operating Rules (DS-AOR) are derived,in order to mitigate the adverse effect of climate change. The objective is to maximize the weighted average annual hydropower generation for all future scenario,and then the Simulation-Based Optimization (SBO) method is implemented to optimize the parameters of DS-AOR. The case study of the Jinxi Reservoir shows that: under uncertain future climate change,DS-AOR is an effective and robust strategy. Compared with Historical Operating Rules (HOR) and adaptive operating rules based on Equal Weights (EW-AOR),DS-AOR results in an increase in hydropower benefits by 0.76*10~8 kWh and 0.61*10~8 kWh,and an increase in hydropower reliability by 0.5%11.17% and 3.50%9.34%,respectively,and it performs more robustly. It is concluded that DS-AOR facilitates adaptive reservoir management under climate change.