During the comprehensive interpretation of geological and seismic data obtained while studying complex reservoirs in West Siberian fields, the methods and principles for processing field geological data to apply neural network-based porosity prediction for reservoirs were examined. The paper presents results of porosity prediction cubes used to identify reservoirs based on processed seismic data, well logging results, and geological information. The work demonstrates an approach to estimating reservoir porosity. Combining well logging interpretation with data on reservoir density variations and using synchronous seismic inversion methods can help in defining a set of probabilistic algorithms for lithotype identification. Based on the application of interactive tools for probabilistic interpretation of well data and 3D seismic surveying, the study provides a prediction and probabilistic assessment of reservoir location with varying saturation levels (water and gas saturation). The results substantiate the presence of oil and gas saturated sandstone within the studied field and horizon with a probability of 50-95 %.