An accurate representation of the climatology of the coupled ocean-atmosphere system in global climate models has strong implications for the reliability of projected climate change inferred by these models. Our previous efforts have identified substantial biases of ocean surface wind stress that are fairly common in two generations of the Coupled Model Intercomparison Project (CMIP) models, relative to QuikSCAT climatology. One of the potential causes of the CMIP model biases is the missing representation of large frozen precipitating hydrometeors (i.e., cloud snow) in all CMIP3 and most CMIP5 models, which has not been investigated previously. We examine the impacts of cloud snow on the radiation and atmospheric circulation, air-sea fluxes, and explore the implications to common biases in CMIP models using the National Center for Atmospheric Research coupled Community Earth System Model (CESM) to perform sensitivity experiments with and without cloud snow radiative effects. This study focuses on the impacts of cloud snow in CESM on ocean surface wind stress and air-sea heat fluxes, as well as their relationship with sea surface temperature (SST) and subsurface ocean temperatures in the Pacific sector. It is found that inclusion of the cloud snow parameterization in CESM reduces the surface wind stress and upper ocean temperature (including SST) biases in the tropical and midlatitude Pacific. The differences in the upper ocean temperature with and without the cloud snow parameterization are consistent with the effect of different strength of vertical mixing due to ocean surface wind stress differences but cannot be explained by the differences in net air-sea heat fluxes.