Aiming at the problems of low recognition accuracy and long processing time of individual fruit information in the fruit sorting process, this study analyzes and forecasts the current fruit sorting production lines at home and abroad. According to the working principle, combining with the industrial robots and visual sorting systems, methods such as the use of convolutional neural network(CNN) and the use of spectroscopic techniques to analyze the chemical properties of fruits are carried out to conduct research on fruit defect detection and maturity sorting. Convolutional neural network and spectral analysis technology in fruit sorting application development trends are described separately. The results show that the research has certain practical value for improving the research of fruit defect detection and sorting accuracy.