SST (Sea Surface Temperature) is one of the most important parameters in the global ocean and climate research. Satellite passive microwave remote sensing is used more and more in SST research to realize all-weather observation. The Microwave Imager (MWRI) on the FY-3 satellite lacks the 7GHz vertical polarization channel, which is more sensitive to SST. In this study, FY-3 MWRI and AQUA AMSR-2 are spatiotemporally matched, and the neural network method is used to simulate the 6.9 GHz vertical polarization brightness temperature (6.9V) of AMSR-2 using the matched MWRI channel brightness temperature. The simulated 6.9V is introduced to the inversion of FY3 MWRI SST. The results show that: After introducing the simulated 6.9V, the inversion accuracy of SST is improved, and the improvement in 35°-90°S is more significant, mainly due to the higher sensitivity of 6.9V to low SST and the less influence of wind speed in the inversion of low SST. If the follow-up satellite of FY-3 can carry 6.9 GHz channel, the retrieval accuracy of the low SST particularly at the higher latitudes will be further improved.