Abstract: In order to realize the rapid nondestructive grade evaluation of Imperial Chrysanthemum, machine vision technology was applied to intelligently grade five grades of Imperial Chrysanthemum in this paper. Firstly, the grading device was designed according to the quality characteristics of Imperial Chrysanthemum, and different grading standards were set according to the color, shape, integrity and other characteristics of Imperial Chrysanthemum. Secondly, the image preprocessing of Imperial Chrysanthemum was completed by using image graying, image denoising and image enhancement technology. Thirdly, RGB model was used to complete the color feature extraction and recognition of Imperial Chrysanthemum, and the image integrity judgment and flower diameter calculation of Imperial Chrysanthemum were completed through image segmentation and edge detection technology, so as to obtain the prediction level of Imperial Chrysanthemum. Finally, a set of Imperial Chrysanthemum intelligent grading system was developed based on Microsoft Visual Studio 2017 platform to realize real-time visual operation. The results showed that the overall classification accuracy of the Imperial Chrysanthemum intelligent grading system designed in this paper reached 97.6%, and the average grading speed was more than 5 times that of manual classification. It was superior to the traditional manual classification in reliability, speed, work efficiency and robustness. This study provided a practical case and technical reference for the application of machine vision technology in the field of scented tea grading.