Abstract: In order to achieve intelligent diagnosis of potato early blight, this article proposes a potato early blight diagnosis method based on "image+spectrum" technology. This system intelligently analyzes the symptoms of potato early blight. Firstly, collect relevant information on potato lesions, and then process the information accordingly, including adjusting colors and cropping images to extract key features of the lesions. Finally, the system utilizes lesion features to identify and automatically diagnose potato early blight. After experimental verification, this method can effectively diagnose potato early blight and provide innovative methods for intelligent monitoring and early warning of plant diseases in smart agriculture.