Efficient management of water and energy resources has become a strategic requirement across integrated mining and phosphate complexes due to rising operating costs, environmental pressures, and sustainability targets. This study proposes an artificial intelligence (AI)-driven framework for optimizing water and energy consumption across the complete phosphate value chain, from phosphate rock mining and beneficiation to slurry transport and downstream processing units including sulfuric acid, phosphoric acid, ammonia utility integration, and di-ammonium phosphate (DAP) production. The framework applies advanced machine learning models, including Long Short-Term Memory (LSTM), Random Forest, and Extreme Gradient Boosting (XGBoost), to predict water demand, energy use, and process efficiency using real-time industrial and operational data. By integrating predictive analytics with optimization techniques, the proposed system supports dynamic decision-making, improved water recycling, reduced energy intensity, and better production coordination across upstream and downstream units. The study further highlights how AI-enabled optimization contributes to lower carbon emissions by reducing excess pumping, thermal losses, and inefficient process loads. The proposed approach supports sustainable industrial development, aligns with digital transformation trends, and provides a scalable pathway for intelligent resource management in integrated phosphate complexes.