One of the main goals of seismic inversion is to obtain high-resolution relative and absolute impedance for reservoir properties prediction. We aim to study whether the results from seismic inversion of subsalt data are sufficiently robust for reliable reservoir characterization. Approximately of poststack, wide-azimuth, anisotropic (vertical transverse isotropic) wave-equation migration seismic data from 50 Outer Continental Shelf blocks in the Green Canyon area of the Gulf of Mexico were inverted in this study. A total of four subsalt wells and four subsalt seismic interpreted horizons were used in the inversion process, and one of the wells was used for a blind test. Our poststack inversion method used an iterative discrete spike inversion method, based on the combination of space-adaptive wavelet processing to invert for relative acoustic impedance. Next, the dips were estimated from seismic data and converted to a horizon-like layer sequence field that was used as one of the inputs into the low-frequency model. The background model was generated by incorporating the well velocities, seismic velocity, seismic interpreted horizons, and the previously derived layer sequence field in the low-frequency model. Then, the relative acoustic impedance volume was scaled by adding the low-frequency model to match the calculated acoustic impedance logs from the wells for absolute acoustic impedance. Finally, the geological information and rock physics data were incorporated into the reservoir properties assessment for sand/shale prediction in two main target reservoirs in the Miocene and Wilcox formations. Overall, the poststack inversion results and the sand/shale prediction showed good ties at the well locations. This was clearly demonstrated in the blind test well. Hence, incorporating rock physics and geology enables poststack inversion in subsalt areas.