In recent years, agricultural modernization process in China has significantly accelerated, and development requirements for agricultural machinery have been put forward to move forward with automation and intelligence．Machine vision technology as core of intelligent agricultural machinery, plays a crucial role in transformation and upgrading process of intelligent agricultural machinery．Current application status of machine vision technology in weed recognition, crop growth information monitoring, agricultural robots was systematically sorted out in research．It was found that in agricultural current application scenarios, machine vision technology faced several challenges, such as algorithm adaptability needing to be improved, model generalization ability being relatively insufficient, low coordination between algorithms and agricultural machinery, and data resources being difficult to obtain and insufficiently diversified．On this basis, future application of machine vision technology in agricultural field was proposed to be in multimodal information fusion, lightweight and efficient recognition algorithms, data augmentation and transfer learning, and algorithms optimization for collaborative adaptability with various types of agricultural machinery．