Plant diseases provide the largest risk to the safety of food. The quantity and quality of agricultural goods might be drastically reduced by them. The main problem facing the agriculture industry is the identification of plant diseases. The automated detection and diagnosis of plant diseases has become a common practice in recent years thanks to the development of machine learning and computer vision techniques. Based on their ability to reliably identify and categorize damaged plants based on their visual characteristics, these strategies have demonstrated encouraging outcomes. Computer vision issues involving picture categorization can be successfully handled by convolutional neural networks (CNN). A web-based tool was created to diagnose plant diseases using faulty leaf images and a suggested model that can both detect the illness and suggest a course of action. The used images dataset includes 6697 leaf photos from three different crops such as tomato, potato, apple, corn, grape and strawberry divided into 37 different classes, including 9 classes for tomato disorders and 1 class for healthy conditions, 2 classes for potato disorders and 1 class for healthy conditions, and 3 class for pepper disorders and 1 class for healthy conditions. The suggested model, according to the findings of our experiment, has the greatest training accuracy of 98.35% and validation accuracy of 94.74%.