Disease detection and prevention is a prevalent factor affecting crop yield and productivity. Deep learning and image processing techniques have become prominent smart farming tools. Therefore, identifying plant diseases from leaf images has been an existing issue in most farming problems. This research focuses on detecting these diseases effectively using deep learning models such as transfer learning based on VGG19 approaches. The proposed VGG19 transfer model has been trained using 42000 images and tested against 43000 unseen samples. This model could classify the disease images and could identify the disease type for the unhealthy specific images. VGG19 with transfer learning has been applied for training the image classification and disease detection based on large datasets of around 60 million images of different kinds of leaves. VGG19 is one of the CNNCNN-based architectures deeplearning neural network Image classification based on VGG model is performed using ReLu and Tanh activation functions. The image dataset consists of unhealthy leaf images of the following leaves apples, corn, potatoes, and grape. The transfer learning neural network model is applied to identify and detect diseases at the early stages of their growth and was proposed. This proposed system utilizes leaf images as the key factor in training the model and later tested with unseen samples for the prediction of diseases to facilitate plant disease detection through image classification techniques. The proposed Leaf Specific VGG model provides a solution for monitoring symptoms of plant leaf disease. The learning rate, activation functions, epochs, batch size, and dropout layers are the set of significant hyperparameters of this model. The proposed VGG model attained the optimal prediction accuracy of 99.76% for the datasets containing 85000 images than other existing methods like SVM, CNN, ImageNet, image processing methods, and VGG16.