By introducing the concept of information content and using Bayesian estimation theory, a scheme for calculating information contents in temperature, humidity, O3, SO2, CH4 and other atmospheric parameters in the remote sensing data is given. The scheme takes into account the channel sensitivity to the atmospheric parameters, the observation error of each channel as well as the error characteristics of the background atmosphere. The temperature and humidity information contents in the observation of HIRS, AMSU, AIRS, IAIS and three combinations of the instruments are calculated using the scheme. The results show that the scheme may have valuable applications in instrument designing, evaluating and comparing different instruments and usage of remote sensing data etc.