Commonly, the combination of rock physics and statistical and probabilistic methods applied to seismic-inversion results has been successful for defining reservoirs. Even so, there are some cases in which it is quite complicated to generate a lithologic discrimination, and predictions cause high uncertainty levels. This is the case for the Orinoco Oil Belt's reservoirs in eastern Venezuela, where lithologic distinction generally is difficult, even with rock-physics analysis. To obtain better discrimination, alternative solutions are required. For this reason, the classification algorithm of support vector machines (SVM) was evaluated. Consequently, the main objective is to perform a lithologic discrimination, generating a lithofacies volume through integration of these three main topics. To reach this goal, a rock-physics analysis was applied over well logs and lithologic information, allowing one to obtain relations between lithologic facies and elastic properties in the reservoirs. Subsequently, through simultaneous seismic inversion, it was possible to generate P- and S-wave impedances and density volumes, using a priori geologic information, well logs, and prestack seismic data as input. Finally, a lithofacies volume was generated using the algorithm of support vector machines as a classification tool. Results of this study allowed the identification of reservoir pay zones, validating the lithofacies volume over blind wells with 70% correlation. Corroboration of the SVM effectiveness as a lithologic classifier showed validations of more than 86% for this particular research.