Crop failure owing to pests & diseases are inherent within Indian agriculture, leading to annual losses of 15 to 25% of productivity, resulting in a huge economic loss. This research analyzes the performance of various optimizers for predictive analysis of plant diseases with deep learning approach. The research uses Convolutional Neural Networks for classification of farm or plant leaf samples of 3 crops into 15 classes. The various optimizers used in this research include RMSprop, Adam and AMSgrad. Optimizers Performance is visualised by plotting the Training and Validation Accuracy and Loss curves, ROC curves and Confusion Matrix. The best performance is achieved using Adam optimizer, with the maximum validation accuracy being 98%. This paper focuses on the research analysis proving that plant diseases can be predicted and pre-empted using deep learning methodology with the help of satellite, drone based or mobile based images that result in reducing crop failure and agricultural losses.