Abstract: Sea surface temperature (SST) plays an important role in climate monitoring, oceanographic research, fisheries economic research and pollution monitoring. In this paper, we investigate the retrieval of sea surface temperature from the Microwave Radiation Imager (MWRI) data of the Fengyun-3D satellite (FY-3D). Firstly, the GPM Microwave Imager (GMI) on board the Global Precipitation Measurement (GPM) Satellite is used as the reference sensor to calibrate the FY-3D MWRI using the double difference method based on the Radiative Transfer Model (RTM), which eliminates the in-orbit radiometric calibration bias and provides a basis for accurate retrieval of the sea surface temperature. Then, based on the waveband selection, a new multi-band retrieval algorithm for sea surface temperature coupled with microwave sea surface emissivity is developed in this paper. Compared with the traditional multi-band retrieval algorithm, the root mean square error of the new retrieval algorithm is reduced from 0.74 K to 0.69 K. Finally, the new retrieval algorithm is used to retrieve the global sea surface temperature from the FY-3D MWRI data in 2020, and the accuracy of the retrieval results is verified with the temporally matched iQuam buoy data and GMI sea surface temperature products, with errors of 0.04±1.13 K and −0.07±1.18 K, respectively. In addition, under the same conditions, the errors of the sea surface temperature obtained by retrieval with the conventional multi-band algorithm are 0.30±1.17 K and 0.10±1.17 K. The theoretical and actual retrieval accuracies of the new algorithm are better than those of the traditional multi-band algorithm, indicating that the new multi-channel retrieval algorithm coupled with microwave sea surface emissivity developed in this paper is effective and accurate.