Abstract: Reinterpreting Professor Ye Duzheng’s concept of a future weather–climate prediction system in the context of global change and rapid advances in Artificial Intelligence (AI) helps clarify both its scientific implications and its contemporary relevance. This paper makes three main points. First, Ye’s ideas can be organized around four linked dimensions: systems organization, uncertainty management, human–nature coupling, and user-oriented service. Second, improvements in prediction skill increasingly depend not on a single model alone, but on a closed-loop system integrating observations, data assimilation, model evolution, verification, and service feedback. Third, AI should be understood not as a substitute for physical models but as an embedded functional module that can enhance multi-source information fusion, probabilistic ensemble generation, bias correction, and surrogate parameterization. Future integrated prediction systems should therefore be built around physical consistency, calibrated uncertainty quantification, and governable update mechanisms.