As a critical link of the Beijing-Taipei high-speed railway, the Min-Tai high-speed railway connects Fuzhou and Taipei, bearing significant strategic importance. The diverse operating environment of the Min-Tai high-speed railway presents considerable challenges for research on automatic driving. Since the Min-Tai high-speed railway has not yet been constructed, real train driving data cannot be obtained. Therefore, this paper utilizes the concept of generative AI to generate a large amount of virtual driving curves and sets standards to select high-quality driving curves. Finally, a deep fuzzy system is used to train the automatic driving model, improving the model's energy efficiency and accuracy, providing forward-looking research for the future automatic driving of the Min-Tai high-speed railway.