Abstract: To address the issue of data silos, data formats and over-reliance on manual interpretation in traditional advanced geological prediction, an intelligent auxiliary decision-making system of the Yangtze-to-Hanjiang River Water Diversion Project is developed. The system integrates data governance, intelligent prediction, and data visualization with key technologies such as knowledge graphs, multidimensional data evaluation, and U-Net neural networks. It establishes an end-to-end technical framework covering data quality assessment, scheme recommendation, data processing, integrated imaging, interpretation, and result evaluation, enabling automated and intelligent geological prediction. The practical application shows that the system is stable, efficient and reliable, and can evaluate and optimize the quality of multi-source geophysical data. The prediction results are consistent with the actual excavation and disclosure, and the expected results are achieved.