<p>In the field of oil and gas exploration engineering, logging data is the key information for obtaining subsurface oil and gas reservoir information. Geophysicists establish accurate formation models through comprehensive logging curve data and then formulate oil and gas development strategies. However, in the actual logging process, due to the fact that instruments are often affected by multiple environmental factors, the formation resistivity change curve is shifted, making it difficult to reflect the real formation resistivity information. Especially for slim hole array lateral logging instruments, which are significantly affected by the borehole, borehole correction processing is urgently needed. To address this problem, this paper combines neural networks with the prediction of borehole correction coefficients for slim hole array lateral logging and proposes a borehole correction coefficient prediction method based on a particle swarm optimization (PSO) and backpropagation (BP) hybrid model. Firstly, this paper uses the traditional BP neural network model to predict the borehole correction coefficient. The results show that the probability that the correction coefficient error is within 5% is 92.3%. To further improve the prediction accuracy of logging curves, this paper uses the PSO-BP neural network model for training and prediction. After verification, the probability that the correction coefficient predicted by the PSO-BP model has an error within 5% is as high as 98.8%. This result indicates that the PSO-BP model has superior and stable performance and can be effectively applied in borehole correction processing. It is an efficient and reliable prediction method for slim hole correction coefficients. The research results of this paper provide strong support for realizing intelligent downhole drilling and have important practical application value.</p>