The study of reservoir permeability and porosity is paramount for effective reservoir management and formulation of a production strategy. The Illizi Basin is a Palaeozoic–Mesozoic intraplate depression that preserves over 7000 m of sedimentary rock record and contains world-class petroleum systems with an estimated ultimate recovery (EUR) of over 39 billion barrels of oil equivalent (BBOE) in hydrocarbon reserves. However, predicting and characterising high-permeability (K) zones in such tight gas reservoirs remains challenging due to their complex geological settings and limited well data. This research addresses the critical dilemma of accurately identifying and classifying high-permeability zones in the Illizi Basin. We propose a novel approach that combines conventional geological, sedimentological, and petrophysical analyses with advanced artificial neural networks (ANNs) optimised using deep learning techniques. The study focuses on the north-western part of the basin, where distinguishing permeability facies using conventional methods is particularly difficult. The novelty of this work lies in the application of a highly efficient ANN model for detecting and classifying high-permeability zones, significantly improving the understanding of permeability distribution within the reservoir. The ANN approach demonstrated exceptional performance, enabling the accurate classification of permeability facies and the detection of high-permeability zones in all wells across the study area. This innovative integration of deep learning with traditional reservoir characterisation techniques provides a more reliable framework for reservoir management in tight gas formations like in the Illizi Basin.