Water vapor plays a key role in the weather, climate, and environment at both local and global scales. Multi-sensory monitoring of water vapor makes it possible to continuously observe water vapor distribution with high spatial and temporal resolutions. Together with remote sensors of MODIS/Terra from the U.S. and MERIS/Envisat from European Space Agency, the Medium Resolution Spectral Imager (MERSI) from the 3 rd generation of Chinese Fengyun satellite is another major sensor for water vapor observation. A simplified retrieval algorithm is proposed for MERSI Near Infrared (NIR) Channels from the Fengyun- 3B satellite. The main feature of the MERSI NIR water vapor retrieval scheme is that it requires no extra information on other atmospheric parameters. The MERSI Level 1b reflectance data are employed in water vapor retrieval. Three absorption channels centered at 905 mm, 940 mm, and 980 mm are used to estimate the transmittance of water vapor. In this work, the water vapor data estimated from ground-based GPS stations in the west North America are used as reference data for MERSI water vapor calibration and validation analysis. GPS PWV is firstly employed to establish the correlation between the reflectance ratio of MERSI and water vapor concentration. Then, validation analysis is performed against the GPS PWV to assess the performance of the proposed algorithm. The calibrated MERSI water vapor data have an RMS value of 5.923 mm, a reduction of nearly 30% compared to the uncalibrated raw water vapor product from MERSI.