Abstract: The Penglai Gas Field is located on the northern slope of the central Sichuan Paleo-Uplift. It has large burial depth, complex pressure systems, and highly abrasive rock formation. The average rate of penetration (ROP) is less than 2.5 m/h; the service life of the bit is only 60–90 hours, and the poor compatibility of the drill bits with the formation seriously restricts the drilling speed improvement. To address this issue, a combined subjective and objective method for polycrystalline diamond compact (PDC) bit selection and drilling parameter optimization was proposed: A drill bit database in the Penglai Gas Field was established, and data preprocessing was conducted. An evaluation method that combined subjective and objective aspects was adopted. By comparing four artificial intelligence models, the CNN-GRU-Attention, which had a higher fitting accuracy, was selected as the prediction model. The optimal drilling parameters combinations were determined through grid search. The results show that after the PDC drill bit selection, the mechanical specific energy is reduced by 28.8% compared to the original. Further optimization of drilling parameters leads to a 31.2% reduction in the mechanical specific energy after selection, providing significant engineering value for improving drilling efficiency in the Penglai Gas Field of Sichuan Basin.