A precision agrochemical application robot installed on elevated guide rails for cucumber disease management in greenhouse was developed. This robot integrated a disease diagnosing and early warning system, a variable agrochemical application system and control system. Diseases were acquired by using machine vision, and quantitative effects were analyzed according to the color and texture for cucumber plants. Relations between environment information cycle data, including illumination, soil temperature, air temperature and moisture, and plant disease valid characteristic were established to realize the early-warning of arising and growing trend of the disease. The environmental elements comparison and lesions on cucumber distribution offered the basis of decision for trajectory planning and target spray. The spray task was performed by a PLC controlled three-degree of freedom Cartesian coordinate manipulator and each individual nozzle. The robot which worked stably and availably was tested in laboratory.