srjecsSRJECSSRJECS2788-9408https://doi.org/10.47310/srjecs.2025.v05i02.009Machine Learning and Conventional Approaches for Reservoir Lithofacies and Petro-Physical Characterization: A Comprehensive ReviewThe petro-physical properties estimation and lithofacies determination are composed of crucial tactics for reservoir perception, as they have a direct effect on the exploration and the production decisions. Conventionally, the mentioned parameters depend on seismic explanation, logging anatomization and core demonstration.  these operations are well committed and functional, such operations are considered ineffectual due to time wasting and struggling to emphasize the natural complicated features of the reservoir. In recent years, improvements in machine learning (ML) revealed a dominant unfamiliar implement allowing computerized lithofacies description with further precise estimation of petro-physical characteristics by enormous sets of data. ML methods improve data processing, modelling detail and efficiency; though, their efficiency still depends on the type of the data and the supervision of geological expertise. By incorporating the power of conventional and ML based models, a more accurate and complete view of the reservoir that can be accomplished. This research exposes how the combination of machine-learning with traditional methods generates a powerful, reciprocal substructure for dependable lithofacies determination and petro-physical property evaluation ultimately improving the general dependability of reservoir description.