Plant leaf disease distresses the growth of the plant. Therefore, early detection and classification of plant leaf disease is vital for healthy crops such as tomatoes. This early detection of tomato leaf disease in plants is also necessary to stop victims in the agriculture field, in tomato leaf disease plants such as bacterial, early blight, septoria, mosaic virus, and one healthy class. Recently, various deep learning-based architectures, such as AlexNet, ResNet50, and VGG16, have been widely applied for classifying and detecting leaf diseases. Primarily, image processing, convolution neural networks, and deep learning methods have been explored to develop a robust detection model to detect, identify, and classify tomato leaf diseases. However, the previous works’ comprehensive analysis of these models is missing. Therefore, this paper analyzes different deep-learning models for Tomato leaf disease detection. It includes data collection, pre-processing, feature extraction, classification, identification, and detection using AlexNet, ResNet50, and VGG16. The experimental analysis is performed on a plant village dataset for a Tomato plant consisting of one plant, four diseases, and one healthy class. The experimental results showed that ResNet yielded the best performance among the used deep learning models, producing an accuracy value of 98.15%, precision of 98.19%, recall of 97.25%, and F1-score of 97.72%. It is followed by the AlexNet model with an accuracy value of 97.8%, precision of 98.45%, recall of 95.82%, and F1-score of 96.14%. The VGG 16 had the lowest accuracy value of 75%, precision of 74.04%, recall of 68.06%, and F1-score of 70.93%.