Abstract. In recent years, the application of automated guided vehicles (AGVs) in the industrial field has been increasing, and the demand for their path planning and tracking has become more and more urgent. This study aims to improve the effectiveness of AGV path planning and path-tracking control and to design a comprehensive hardware and software system in combination with the Robot Operating System (ROS) to improve the practicality of the system. First, the real-time performance and accuracy of path planning by optimizing window size and dynamic adjustment strategies are improved. Secondly, the research on the fusion of the improved particle swarm algorithm and PID (proportional, integral, differential) control applied to path tracking is discussed in depth. By combining the two organically, the accuracy and robustness of AGV path tracking in complex environments are improved. In the hardware and software system design phase, the ROS provides a more flexible and modular solution for the AGV system, and the introduction of ROS not only simplifies the integration of system components, but also provides a convenient framework for future system upgrades and expansions. In the experimental phase, the methodology adopted in the study is described in detail, and the superior performance of the improved method over the traditional method is demonstrated. The experimental results not only confirm the effectiveness of the improved method in improving path planning and path-tracking accuracy, but also provide strong support for the active role of the ROS in AGV system design.