Agriculture industry can be significantly affected by pests. Deep learning techniques are very useful for detecting various diseases in crops. A dataset consisting of 1530 images was divided into the training set, testing set, and validation set. It consists of labels such as “Healthy”, “Powdery,” and “Rusty.” These labels reflect the conditions of a plant and are used to train the model. The trained models using Deep Convolutional Neural Networks successfully find the relevant crop diseases with a 95% accuracy rate. The early detection of diseases is useful to curb the diseases in the budding stage to avoid the rampant onset of plant diseases which prove fatal to the agricultural field. By implementing this strategy, crop loss can be reduced significantly. In this paper, a detailed explanation of various CNN architectures such as Xception, VGG16, Resnet50, DenseNet, efficientNet, and InceptionV3 is outlined. Each model is trained with training data to predict a plant’s disease. The model’s performance is highly dependent on extracting features from the data. The result analysis and visualizations of obtained results are presented in the analysis section of the paper.