Abstract: This paper introduces the application of ocean temperature and salinity observations from the Global Telecommunication System (GTS) in the new ocean data assimilation system of the National Climate Center. Using quality control on temperature and salinity data, we judge the duplicate records, observation depth, topography, extreme value, climate variability, stratification and spatial difference, so that false and trustless observations are effectively filtered. After quality control, the temperature and salinity data are added into the assimilation system, which effectively improves the MOM4’s simulation on the distribution features of the multiyear sea surface temperature and salinity in global tropics and subtropics, especially in the Pacific Ocean. In addition, the assimilation also has an obvious improvement to the distribution features of sea surface temperatures in the middle latitudes of both hemispheres. The comparison of the root mean square error (RMSE) between with and without assimilation shows that, after assimilation, the RMSEs of the sea surface temperature and salinity in most ocean regions, especially in tropic regions, have a significant decline with a range of 0.1-1.0, therefore the simulated distribution features of sea surface temperature and salinity are closer to the observations. The further analysis indicates that the assimilation clearly improves the simulation on the time variation features of Nino3 and Nino4 sea surface temperatures. For the Nino3 temperature, its difference with the OISST is decreased after assimilation, always having a small absolute value of difference of about 0.5℃ in first halfyear and a large one of about 1℃ in second halfyear. For the Nino4 temperature, it is improved obviously and the absolute value of its difference with the OISST is always below 0.5℃.