Seismic inversion associates uncertainties due to the band-limited nature of the seismic data. Deterministic methods of seismic inversion are not appropriate for reservoir characterization because of the inability to model the heterogeneities and uncertainties. Stochastic seismic inversion, in addition to model heterogeneities, provides the possibility to calculate uncertainties in the reservoir. In this research, it is assumed that the probability density functions of the logarithm of acoustic and shear impedances are Gaussian and these functions were obtained using Bayesian principle. The sampling of the global functions was performed using sequential Gaussian simulation, and multiple realizations of acoustic and shear impedances were generated. Applying this method to one of the Persian Gulf oilfields showed that the stochastic inversion is capable of modeling the heterogeneities and recovers the high-resolution thin beds.