Sea Surface Temperature (SST) is one of the important parameters of hydrodynamic and oceanographic research. It has been widely utilized in the study of climate change, weather forecasting, numerical weather prediction(NWP), atmospheric and oceanographic applications, fisheries, and other sciences. This study comparesSST derived from WindSat to those directly observed by NDBC/TAO buoy and SST provided by AMSR-E. In the collocation of WindSat and buoy data, an SST value with a rain rate larger than zero was rejected. Collocation adopted a 30 min time constraint and 25 km maximum separation. A total of 82,247 WindSat-NDBC collocations and 366,693 Wind-Sat-TAO collocations were acquired. Rsults show that WindSat SST has a mean bias of 0.10 ℃ and standard deviation of 0.59 ℃ in the coastal and offshore waters of the United States. In the tropical Pacific Ocean, WindSat SST has a mean bias of-0.15 ℃ and standard deviation of 0.33 ℃. In addition, WindSat SST has an increasing warm bias or a decreasing cool bias in summer. The standard deviations of the satellite-derived SST are both relatively large in the US East Coast and in the Gulf of Mexico; the standard deviations of certain regions are higher than 1 ℃. When the buoy wind speed ranges from 5 m/s to 10 m/s, the accuracy of WindSat SST is good, and the mean bias and standard deviation are relatively constant. When the buoy wind speed exceeds 12m/s, the uncertainty of WindSat SST increases. Global comparison with AMSR-E shows that WindSat monthly averaged SST is cooler than that of AMSR-Ein summer and warmer in winter. Even though it has operated beyond its designed life of seven years, WindSat SST continues to exhibit acceptable accuracy. WindSat retrieves SST better in tropical Pacific and worse in US coasts and in the Gulf of Mexico. The part of wind direction correction in retrieval algorithms needs to be improved to increase the accuracy of SST retrieval under conditions of high wind speed.