Navigation system design plays an important role in modern agricultural machinery navigation, but there is a problem of inaccurate navigation and positioning. The traditional particle swarm algorithm cannot solve the system design problems in modern agricultural machinery navigation, and the effect is not ideal. With the rapid development of science and technology, the field of agriculture has also ushered in unprecedented changes. Traditional farming methods have been gradually replaced by intelligent and automated technologies. In this revolution of agricultural science and technology, the application of the Extreme Learning Machine (ELM) image intelligent classification algorithm is particularly eye-catching, which brings subversive improvements to the modern agricultural machinery navigation system.