Many global climate models are prone to producing land-atmosphere coupling dynamics that are too strong and simplistic. Cumulus and convection parameterizations are natural culprits but the effect of bypassing them with explicitly resolved convection on global land-atmosphere coupling dynamics has not been explored systematically. We apply a suite of modern land-atmosphere coupling diagnostics to isolate the effect of cloud superparameterization (SP) in the Community Atmosphere Model v3.5, focusing on both the land segment (i.e., soil moisture and evapotranspiration relationship) and atmospheric segment (i.e., evapotranspiration and precipitation relationship) in the water pathway of the land-atmosphere feedback loop. Comparing SPCAM3.5 and conventional CAM3.5 in daily timescale, our results show that the Super-Parameterized model reduces the coupling strength in the Central Great Plain in American, and reverses the terrestrial segment coupling sign (from negative to positive) over India. Which are consistent with previous studies and are favorable improvements on the known issues reported in literatures. Analysis of the triggering feedback strength (TFS) and amount feedback strength (AFS) shows that SPCAM3.5 favorably reproduces the patterns of these indices over North America, with probability of afternoon precipitation enhanced by high evaporative fraction along the eastern United States and Mexico, while conventional CAM3.5 does not capture this signal. The links in the soil moisture-precipitation feedback loop are further explored through applying the mixing diagram approach to the diurnal cycles of the land surface and planetary boundary layer variables.