This paper presents the development of an algorithm to retrieve the canopy water content of natural grasslands and pasture from synthetic aperture radar (SAR) measurements. The development of this algorithm involves three interrelated steps: (1) calibration of SAR data for ground topographic variations, (2) development and validation of backscatter model for grass canopies, and (3) estimation of canopy water content by inverting a backscattered model for cross-polarized ratio. The polarimetric radar data acquired by the Jet Propulsion Laboratory AIRSAR system during the 1989 First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE) are used for this study. The SAR data have been calibrated and corrected for the topographical effects by using the digital elevation map of the study area. The backscattering coefficients obtained from the SAR data for each pixel are related to the canopy and soil parameters by employing a discrete random media model for vegetation. The model simulations indicate that biomass variations and surface treatments (burned and unburned) of grass canopies affect the C-band backscatter signal but does not influence the L-band signal. This model is then validated and adjusted over training areas where ground measurements were collected. An inversion technique is proposed to estimate the canopy water content by using the cross-polarized and copolarized ratios of the SAR data at C band. The result of the inversion algorithm shows a good agreement with the grass biomass data collected during FIFE 1989 intensive field campaign. ¿ American Geophysical Union 1995