Abstract: The detection of cloud liquid water (CLW) is very important in current meteorological service and research. The precipitation process is influenced by the interaction between CLW, vapor and ice phase particles, so the changes of global climate and local weather are influenced by the distribution of CLW. The detection of CLW can be used to identify the artificial precipitation enhancement potential regions, so the efficiency of weather modification can be improved. The liable data needed in numerical weather prediction can be obtained by the detection of CLW, and the study of numerical prediction model can be validated by the observational results of CLW. Now, the application of satellite detection can be used for monitoring the large scale and whole process of the CLW.The TRMM Microwave Imager (TMI) 85.5 GHz channel vertical polarization brightness temperature and the vertical atmospheric properties (including temperature, pressure and humidity) contained in 4 times daily 1°×1°NCEP data are used to calculate the surface emissivity in Henan Province (31.4°—36.0°N, 110.4°—116.0°E) on March 6, 2005, by means of a numerical step-by-step method with VDISORT. The surface emissivity is supposed to be unaltered in short time. The emissivity of test area on March 21 is considered to be the same with the emissivity on March 21. The cloud top height is calculated with the TRMM/VIRS 12.0 μm channel infrared radiation data, and the cloud bottom height is supposed to be the local lifting condensation level calculated with the terperature and dew point temperature. The retrieved cloud is suppose to be vertically uniform distribution, then the CLW is retrieved by means of iteration method with TRMM Microwave Imager (TMI) 85.5 GHz channel vertical polarization brightness temperature and NCEP data. The calculated surface emissivity is evaluated with the Henan map, and the retrieved CLW is compared with the CLW data from TRMM 2A12 products in the same time and the CLW data from NCEP data, infrared cloud picture in the same period. The retrieved CLW is coincident with the distribution of cloud in infrared cloud picture, and shows improvements comparing with the CLW data from TRMM 2A12 products, from which the data are probably related to the precipitation, so the range of CLW is more smaller than retrieved CLW and the valid data of CLW are too few. The CLW data from NCEP are not corresponding to either the retrieved CLW or the distribution of cloud in infrared cloud picture. The model simulation results show that the method is more accurate in retrieving high cloud CLW than low cloud. The error is increased by the use of NCEP data instead of sounding data in the CLW retrieving. So in the future the condensed sounding data should be used in CLW retrivel test. The comparative analysis is mostly qualitative for the lack of measurements, so the quantitative comparative analysis should be done more in the future study.