The development degree of shale laminations directly affects the productivity of shale oil reservoirs. However, current research on shale laminations mainly focuses on qualitative studies such as lamination types, lacking quantitative identification methods. This paper aims to propose an intelligent method for quantifying shale lamination numbers. By utilizing the Classification and Regression Tree (C4.5) algorithm based on the Gini coefficient, coupled with grid search optimization, parameters including static resistivity, lithology, natural gamma, acoustic travel time, compensated neutron, and density from imaging logging are employed as inputs. Shale laminations are calibrated using two different resolution images, core samples, and imaging logging. The model is then trained and optimized continuously, yielding the following results: the optimal parameters for grid search optimization are a tree depth of 6, minimum samples per leaf of 5, and minimum samples per split of 25. After optimization, the prediction accuracy of the C4.5 model increases from 78 to 85.3%, with a cross-validation accuracy of 81%. In practical applications, a lamination rate of 31.4% is calculated for a section of well C. The refined C4.5 model establishes a mapping method between imaging logging and conventional logging, providing a reliable solution for identifying shale laminations using conventional logging while also supporting research on sweet spot selection. The innovation of this study lies in transforming the complex problem of lamination identification into a binary classification problem, combined with the C4.5 algorithm based on the Gini coefficient to obtain clear lamination classification criteria, thereby making the identification of lamination numbers more intuitive and accurate.