This document describes a system for diagnosing crop leaf diseases using convolutional neural networks. The system can identify diseases in five major crops: corn, rice, wheat, sugarcane, and grapes. It uses a MobileNet model and CNN architecture trained on datasets of images of healthy and diseased leaves. The system achieves 97.33% accuracy in diagnosing diseases in grape leaves. It aims to help farmers detect diseases early and determine the appropriate pesticides.