With an aging rural population and declining birth rates posing challenges to agricultural productivity, we propose an AI Agriculture Automated Guided Vehicles system (A3GV). It leverages machine learning for real-time image recognition, enabling the vehicle to follow a target. ROS serves as the software framework for sensor coordination, offering features like Simultaneous Localization And Mapping (SLAM), obstacle avoidance, and autonomous navigation. This system alleviates the agricultural workload, and its adaptability is enhanced through a collaborative Unmanned Aerial Vehicle (UAV) setup. A dedicated mobile app enhances user experience by allowing remote vehicle control, mode switching, and autonomous navigation based on specific scenarios.