Channel selection experiments for the retrieval of temperature have been carried out using the data obtained by GIIRS (the hyperspectral vertical detector equipped on the FY-4A geostationary satellite). According to information capacity, using line-by-line radiative transfer model, channels are selected by step-by-step iterative method. Finally, sensitive channels using for retrieval of temperature profile are selected, 56 channels in global and 58 in FY-4A’s detection range. The selected channels have small correlation with each other, and they have higher information capacity, which can be used as the representative channel set for temperature profile retrieval research and data assimilation.