There are many approaches and concepts for the exploration and development of the hydrocarbon reservoirs. In this study, the aim is predicting the gas volume which is a multiplication of the effective porosity with
hydrocarbon saturation. Predicting the gas volume away from the wells is a challenging task because of the nonuniqueness
in its relationship with the conventional seismic attributes. A conjunction of a set of seismic attributes
obtained from multi-linear regression technique with the non-linearity of artificial neural networks techniques can be
utilized to develop effective workflows to explore the reservoirs and evaluate hydrocarbon presence. Insufficient wells
in the studied area led us to develop lithology classification workflow to increase the reliability of predicting gas
volume probability cube over the studied area. Within the studied area, which covers around 660 square km, reservoirs
are mainly Pliocene slope channel system and consist of a succession of sandstones and mudstones organized into a
composite upward finning profile. The matching in the presence of gas volume in the sand classes comes up with a
possibility of prospect evaluation at each location inside the 3D seismic coverage. Results suggest that the application
of the proposed neural network method leads to reliable inferences and has a positive impact on the exploration or
development over the area of study