It has long been recognized that differences in climate model-simulated cloud feedbacks are a primary source of uncertainties for the model-predicted surface temperature change induced by increasing greenhouse gases such as CO2. Large-scale circulation broadly determines when and where clouds form and how they evolve. However, the linkage between large-scale circulation change and cloud radiative effect (CRE) change under global warming has not been thoroughly studied. By analyzing 15 climate models, we show that the change of the Hadley Circulation exhibits meridionally varying weakening and strengthening structures, physically consistent with the cloud changes in distinct cloud regimes. The regions that experience a weakening (strengthening) of the zonal-mean circulation account for 54% (46%) of the multimodel-mean top-of-atmosphere (TOA) CRE change integrated over 45°S–40°N. The simulated Hadley Circulation structure changes per degree of surface warming differ greatly between the models, and the intermodel spread in the Hadley Circulation change is well correlated with the intermodel spread in the TOA CRE change. This correlation underscores the close interactions between large-scale circulation and clouds and suggests that the uncertainties of cloud feedbacks and climate sensitivity reside in the intimate coupling between large-scale circulation and clouds. New model performance metrics proposed in this work, which emphasize how models reproduce satellite-observed spatial variations of zonal-mean cloud fraction and relative humidity associated with the Hadley Circulation, indicate that the models closer to the satellite observations tend to have equilibrium climate sensitivity higher than the multimodel mean.