This paper presents an autonomous multi-trolley collection system with nonholonomic robots, designed to efficiently collect and transport multiple luggage trolleys in dynamic environments like airports. The system integrates a lightweight manipulator and docking mechanism, optimized for sequential stacking and transportation. A novel vision-based control strategy, based on Control Lyapunov Function (CLF) and Control Barrier Function (CBF), is proposed to ensure accurate and efficient collection. The system is tested in real-world scenarios, demonstrating successful execution of multiple-trolley collection tasks. The key contributions include a cost-effective hardware design, robust perception, dynamic motion planning, and optimization-based control. The system is capable of navigating and manipulating trolleys in a sequential manner, decomposing the task into four stages: Searching, Approaching, Docking, and Transportation. The vision-based controller ensures the Field-of-View (FoV) constraint and high tracking accuracy. The system is demonstrated in a real-world scenario, where it successfully collects multiple luggage trolleys. The experiments show that the proposed controller improves the success rate of the multiple-trolley collection task. The system is compared with other control strategies, showing superior performance in terms of control accuracy and robustness. The results demonstrate the effectiveness of the proposed system in real-world applications.This paper presents an autonomous multi-trolley collection system with nonholonomic robots, designed to efficiently collect and transport multiple luggage trolleys in dynamic environments like airports. The system integrates a lightweight manipulator and docking mechanism, optimized for sequential stacking and transportation. A novel vision-based control strategy, based on Control Lyapunov Function (CLF) and Control Barrier Function (CBF), is proposed to ensure accurate and efficient collection. The system is tested in real-world scenarios, demonstrating successful execution of multiple-trolley collection tasks. The key contributions include a cost-effective hardware design, robust perception, dynamic motion planning, and optimization-based control. The system is capable of navigating and manipulating trolleys in a sequential manner, decomposing the task into four stages: Searching, Approaching, Docking, and Transportation. The vision-based controller ensures the Field-of-View (FoV) constraint and high tracking accuracy. The system is demonstrated in a real-world scenario, where it successfully collects multiple luggage trolleys. The experiments show that the proposed controller improves the success rate of the multiple-trolley collection task. The system is compared with other control strategies, showing superior performance in terms of control accuracy and robustness. The results demonstrate the effectiveness of the proposed system in real-world applications.