<p>This study presents a method to correct the lithology of mud-logging profile with logging data based on neural network, which aims to solve the problems of time-consuming, high labor intensity and great influence of human factors in the process of traditional lithology correction of mud-logging profile. Firstly, the lithology of mud-logging profile is processed by digital technology and converted into digital curve which is consistent with the logging sampling interval, and the logging lithology curve is calculated by using the optimal logging method. Then, combining automatic depth-correction technology with manual correction methods, the lithology of mud-logging profile is corrected for depth. On the basis of lithology depth-correction of mud-logging profile, the multi-layer perceptron (MLP) neural network is used to learn logging data and realize accurate identification of multiple lithologies, so as to construct a high-precision logging profile lithology curve and provide accurate basis for lithology correction of mud-logging profile. The effectiveness and accuracy of the proposed method are verified by practical application cases. The corrected lithology of mud-logging profile is highly consistent with the lithology of logging profile, which provides a solid foundation for subsequent geological interpretation, reservoir evaluation and oil and gas resource assessment. This study not only improves the efficiency of mud-logging data processing, but also ensures that the needs of exploration and exploitation work are met in a timely manner, which has important theoretical significance and application value.</p>