To address the scheduling requirements arising from the deep integration of logistics and production in intelligent manufacturing environments, this study proposes a collaborative optimization approach that integrates automated guided vehicle (AGV) scheduling with flexible job shop scheduling. A mixed-integer linear programming (MILP) model is developed to jointly consider workpiece processing path selection, machine tool allocation, and AGV transportation task scheduling. The model incorporates key practical factors such as transportation time, AGV quantity constraints, operation sequence dependencies, and resource availability. To tackle the high complexity of problem-solving, a hybrid intelligent algorithm combining a genetic algorithm and an improved variable neighborhood search (GAIVNS) strategy is designed to enhance solution accuracy and stability. Simulation experiments conducted in a new energy vehicle assembly scenario demonstrate that the proposed model and algorithm outperform existing methods in terms of task completion time, equipment utilization, and scheduling stability. The results indicate that this research provides an efficient and practical solution for multi-resource collaborative scheduling in intelligent manufacturing systems, showing strong potential for engineering applications.