Plant disease is seen as a significant danger to crop productivity in agricultural regions. Plant diseases are often divided into three categories: fungal, viral, and bacterial. Crops can be protected against disease severity by early identification and detection. Farmers used to have a very hard time diagnosing illness with their unaided eyes. A class of deep learning algorithms that automatically learns and extracts features for plant disease identification and detection, including convolutional neural networks. The plant village data set that was used in this instance is openly accessible. There are 20638 leaf images in the data collection, organized into 15 classes. That is, training accounted for 80% of the data set, whereas testing accounted for 20%. The model’s performance is enhanced and the loss function is minimized by using Adam optimizers. Various pre-trained models, including ResNet 101, VGG16, and Mobile Net V2, were compared. It is clear from the experiment’s results that ResNet 101 can classify objects with up to 97.23% accuracy while requiring the fewest number of computations.