Tomato is one of the most cultivated and edible crops in the world. Depending on how fertilized, tomatoes range in number. The key factor determining the volume and cost of crop output is diseased leaves. As a result, it is vital to recognize and classify these root causes of the disease illnesses accurately. Different diseases affect tomato production. The detection of these illnesses at an early stage would reduce the disease’s impact on tomato plants and improve crop production. Several new approaches to specific diseases have been employed widely. This research investigates the efficacy of Convolutional Neural Networks (CNNs) in identifying tomato leaf diseases, crucial for the agricultural economy and food security. The work involves creating a dataset comprising healthy and diseased tomato leaves, including common ailments like bacterial spots, late blight, septoria, mosaic virus, and early blight. Preprocessing methods enhance the dataset quality of the tomato image samples. A CNN architecture is then trained to classify tomato leaves as healthy or diseased. The metrics like accuracy, precision, recall, and F1-score gauge model effectiveness indicate the performance. The model shows the different-accuracy ranging from 99.00%–95.00% for different diseases. Comparative analysis with conventional methods and other deep learning architectures validates CNN’s superiority. Experimental results showcase that CNN-based strategy outperforms traditional methods and yields competitive outcomes compared to alternative deep learning architectures. In order to assist farmers in reliably diagnosing illnesses and alerting them at an early stage about the disease is the purpose of this paper. Transfer learning techniques are explored to enhance disease diagnosis with limited labeled data, leveraging pre-trained CNN models. This study underscores the transformative potential of CNNs in plant disease management, offering scalable solutions for safeguarding agriculture output and food security.