In this study, we present the design and implementation of a batch recycling and counting system for acupuncture needles, addressing safety and operational challenges in traditional Chinese medicine practices. The system incorporates sensors and relevant technologies, including infrared sensors for detecting tray or hand proximity, a high-frequency ultrasonic vibrator for needle dispersion, and an LED-illuminated imaging system for precise recognition using AI-based image processing. The closed imaging space minimizes light interference, enhancing the reliability of the image recognition process. The hardware architecture integrates these sensors with a Raspberry Pi 4 for data processing and control, while network communication employs Restful application programming interface technology for real-time data synchronization and traceability. Laboratory tests achieved a recognition accuracy of more than 95% for batches containing up to 15 needles, with a decline to 87% for batches of 25 needles due to tray saturation. Field verification on 104 patients and 552 needles demonstrated an accuracy of 100% in identifying and matching needles, with an average of 5.3 needles per session. Feedback from healthcare professionals confirmed the system’s practicality, while patient interviews revealed improved confidence in medical safety. Despite some connectivity issues, this system demonstrates the potential of sensor-based technologies in enhancing workflow efficiency and safety in medical waste management.