This paper proposes a two-step approach for characterizing the reservoir properties of the world’s largest non-associated gas reservoir. This approach integrates geological and petrophysical data and compares them with the field performance analysis to achieve a practical electrofacies clustering. Porosity and permeability prediction is done on the basis of linear functions, succeeding the electrofacies clustering. At the start, an unsupervised neural network was employed based on the self-organizing map (SOM) technique to identify and extract electrofacies groups. No subdivision of the data set was required for the technique on account of the natural characters of the well logs that reflect lithological character of the formations. The second step was examining a supervised neural network which is designed based on the back propagation algorithm. This technique quantitatively predicts the porosity and permeability within the determined electrofacies. The final part of the study was calibration and comparison of the electrofacies clustering results with core and petrographic data. Based on the porosity and permeability maps at different depth levels, the target reservoir is classified into six electrofacies clusters (EF1-EF6) among which the EF5 and EF4 show the best reservoir quality.