Objective: In order to solve the problems on tea production of low efficiency of manual sorting, high cost of equipment, large volume, single function and so on. Methods: A small intelligent tea sorting control system was developed by using image recognition technology. According to the gravity, volume, color and shape characteristics of tea and impurities, the system can adjust the wind speed, vibration frequency, color and shape characteristic parameters, and automatically separates tea and impurities. Taking Curl, grain, flake machine-picked tea as the experimental objects, the system performance was tested. Results: The average recognition rate of the control system for the three kinds of machine-picked tea were 86.03%, 93.53% and 89.46% respectively. The yield was about 45.8 kg/h, with the picking rate of about 97%, and the takeout ratio of about 5.4∶1. Conclusion: The intelligent tea sorting control system effectively improves the quality and value of tea and enhances the sorting efficiency.