Agriculture is the cornerstone of human society and the core of food supply, playing an irreplaceable role in global food demand. In order to meet the social demand for high-yield and high-quality crops, we must deal with the potential threats from various diseases and insect pests, such as corn, wheat and some vegetables, which are vulnerable to various diseases and insect pests in the seedling stage, which will greatly reduce their yield and quality. In order to effectively prevent the threat of these diseases and insect pests, agricultural production usually adopts plant protection unmanned aerial vehicle monitoring, insect situation detection lights and manual pest search methods to deal with them. However, these schemes have the disadvantages of high price, low popularization rate and high labor cost. In order to solve these problems and help increase agricultural production and quality, we have designed a wheel-legged agricultural disease and insect pest detection robot based on YOLO V5, OmniMaster. The robot deploys NVIDIA Jetson NX high computing power platform, uses advanced deep learning and target detection technology to provide a variety of detection modes, which can realize fast, accurate and real-time pest detection. At the same time, the robot adopts a mechanical structure design combining wheel type and leg type, is equipped with terrain perception technology, and has high adaptability to complex terrain with excellent electric control system. The robot is also equipped with multi-sensor fusion analysis technology. By detecting crop diseases and insect pests, combined with sensor data, it can comprehensively model and evaluate crop growth and health index, provide farmers with corresponding scientific management measures, significantly reduce the burden of farmers and improve Agricultural production efficiency. This project provides new ideas and directions for the realization of agricultural automation and intelligence, and has a positive role in promoting agricultural modernization.