The precise identification and classification of crop diseases pose a significant challenge in contemporary agriculture, impacting yield, quality, and food security. Traditional diagnostic techniques often prove inadequate due to their dependence on manual effort and subjective judgment— limitations that are particularly evident in large-scale farming systems. This article investigates how deep learning models can be utilized for classifying crop diseases, presenting a more efficient and scalable solution to disease diagnosis challenges. Utilizing advanced deep learning methodologies such as convolutional neural networks (CNNs) and transfer learning, the proposed model is trained with comprehensive datasets containing plant images to accurately identify diverse crop diseases. This study illustrates the application of these models across various crops like maize, cotton, and soyabean while addressing both widespread and uncommon illnesses. A dataset comprising 4,188 leaf images was collected from various agricultural fields under different environmental conditions. The images were categorized into four distinct classes, representing both healthy and diseased leaves. The experimental outcomes reveal marked improvements in classification accuracy over traditional methods; this leads to increasingly timely decisions regarding crop management strategies. model demonstrated excellent performance, achieving an average training and validation accuracy of 100%. To comprehensively evaluate the model, additional performance metrics such as precision, recall, F1-score, and the confusion matrix were analysed