Plants are important and essential elements for humans. Automatic detection and analysis of plant diseases are vital for conserving resources, minimizing yield losses, and enhancing treatment efficacy. This has resulted in healthier plants and more efficient farming practices. A novel automated system was developed to accurately identify and classify diseases using plant images. Using data from plantvillages, this system demonstrates computer vision techniques, such as image processing, machine learning (ML), and deep learning (DL), to go beyond conventional approaches to disease management for important plant diseases such as bacterial spot of tomato, tomato early blight. Diseases were then identified based on visual characteristics using lightweight models (Xception, EfficientNet, VGG16, ResNet50, NASNet Mobile, and MobileNet). The DL approach, utilizing ReLU and softmax functions, achieved an impressive validation accuracy of 99.39%. After identifying the diseases, the system suggests predictive treatments to help farmers and agricultural organizations take effective measures against these threats.