Machine learning and artificial intelligence have been more essential in recent years in areas of agriculture including the diagnosis and forecasting of plant diseases. Because symptoms, field crops, and climatic conditions all vary, it can be difficult to identify plant diseases early on. The quality and quantity of potatoes are affected by a number of diseases, including late blight and early blight. The manual diagnosis of potato leaves disease is a laborious and complicated operation. To diagnose the condition, a specialist with good abilities is needed. Therefore, a technique that can identify the illnesses affecting potato leaves must be automated and effective. In order to extract the deep features from the dataset, a unique convolutional neural network model is developed in this study. On the Plant Village dataset and the Plant Leaf Disease dataset, the model is assessed. The outcome demonstrates that the suggested model outperforms the earlier work in terms of efficiency and outcomes.