Abstract: Vision-based detection of agricultural pests is an important technology for achieving automated and real-time monitoring of pest conditions. Firstly, this paper introduces the application of traditional machine learning techniques in the vision-based detection of pests in China and internationally. Then, this paper summarizes the research progress of the new generation of vision-based detection methods for pests, which are based on deep learning techniques such as R-CNN, Fast R-CNN, Faster R-CNN, SSD, and YOLO. Next, this paper analyzes the problems that exist in the research and practical applications of vision-based detection methods for agricultural pests. The traditional machine learning-based methods have problems such as low feature capture ability, detection accuracy, and robustness, as well as high resource consumption. The deep learning-based methods have higher detection performance than the traditional machine learning-based methods but have problems such as poor performance on small and differently distributed targets, low detection accuracy, and slow speed. Finally, this paper discusses possible research directions in the future for the vision-based detection of agricultural pests based on deep learning techniques, including the development of public resources for agricultural pest image data, robust handling for data distribution shifts, deep feature learning, and multi-scene applications.