This paper evaluates the simulations of East Asian Summer Monsoon (EASM) for the new version of climate system model, namely FGOALS_s, developed at the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics, Institute of Atmospheric Physics (LASG/IAP), which employs the recently improved version of LASG/IAP spectral AGCM SAMIL as its atmospheric component. To understand the impacts of air－sea coupling, the results of coupled system are then compared to AMIP-type SST-forced simulations with SAMIL. FGOALS_s is shown to capture the main climate mean states and seasonal features of EASM, though the bias is obvious. The main deficiency of the FGOALS_s model is that the simulated temperature is systematically colder than the observation in the middle and upper troposphere, which notably weaken the atmosphere circulations and winds. In addition, the insufficiency of meridional air temperature gradient directly influences the simulation of subtropic westerly jet. The comparison between FGOALS and SAMIL demonstrates that there are similar biases in the stand-alone AGCM, though the simulation of SAMIL is more realistic. This indicates that the biases included by the atmosphere model have great influence on the coupling system performance. Besides, the spatial pattern of EASM precipitation is more reasonable in the coupled model, which illuminates that air－sea interaction can improve the simulation of precipitation in EASM while amplify the biases in the temperature field. Therefore air-sea coupling also plays a vital role in the performance of the climate system model. The results suggest that the biases in the atmosphere are the key reasons for the insufficiency of coupled system. From this point, updating cloud and radiation processes to improve the simulation of temperature in the atmosphere model, may be the first step for the future development of FGOALS_s.