Sea Surface Temperature (SST) has an important application value in the field of fishery environment and is also a key environmental element in scientific survey. In this paper, the thermal infrared band of Landsat-5 TM was used to retrieve SST with three different methods, i.e., QIN et al.’s mono-window algorithm, Jiménez-Munn~oz single-channel method and Weng’s algorithm. The first two are simplified models that only consider the atmospheric parameter of water vapor content, QIN et al.’s model contains the atmospheric temperature of the average value as well, but Weng’s algorithm has no elimination of atmospheric parameter. The retrieved data were mapped to analyze the distribution of SST in the Daya bay, and the results are compared with data from 12 sampling points. The results show that distribution characteristics of SST obtained from the three methods are similar to each other, and the diffusion of warm water from the Daya and the Lingao nuclear power stations can be observed clearly. The three methods show an error of -2.21℃, 0.19℃ and -4.68℃ respectively, with the error of the Jiménez-Munn~oz single-channel method being the lowest.