The feeding model and control system of the industrialized South American white shrimp farming system are disconnected, and a closed-loop feeding decision control system has not yet been formed, resulting in insufficient intelligence level of the integrated management and control system for aquaculture production. Therefore, a shrimp feeding dynamic decision-making system based on multi-source information fusion has been developed. The system integrates underwater cameras and dedicated feeding platforms to construct a synchronized monitoring system for shrimp body length and residual bait amount; On this basis, a dual factor feeding decision model that integrates growth status and real-time residual feeding feedback is proposed, endowing the system with adaptive optimization capability; Further design a control execution unit with PLC and upper computer as the core to drive the high-precision feeding machine to achieve reliable implementation of strategies and data management. Comparative experiments have shown that the system significantly improves feeding control accuracy and feed utilization efficiency, increasing shrimp body mass growth rate by 0.26%, reducing feed coefficient by 0.14%, and increasing survival rate by 0.9%. This provides a reliable technical solution for the refinement, intelligence, and sustainable development of shrimp farming.