Based on the atmospheric radiation transmission theory,VIS(visible light)and IR(infrared)spectral characteristics of clouds and fog in the meteorological satellite remote-sensing images are studied in terms of spectral analysis method, which describes the spectrum response features and related index in the each channel of NOAA. According to the different cloud and fog top structure,the non-lambert property of cloud/fog reflecting visible light are analyzed.Based on the Rayleigh criterion,it is explained that visible light possesses more abilities of identifying fog and cloud when sun’s altitude is lower,indicating that fog has stronger directivity reflecting characteristics than cloud.When the satellite is placed in the mirror-reflecting-direction of cloud and fog,the brightness is stronger in fog area than in cloud area,otherwise darker.This shows that visible light has important role in identifying fog and cloud.By analyzing the different IR(infrared)spectral characteristics of cloud and fog layers,some indices of identifying cloud and fog and their classifications are given, showing that the channel combination method is of significance in identifying cloud and fog, and provides proofs for the classification of cloud and fog.By analyzing the time-variation spectrum of fog,the spectral characteristics of heavy fog in mature phase are obtained,which has vital meaning on judging the development and dissipation of heavy fog.