Early and correct detection of plant diseases is crucial for maintaining agricultural productivity and reducing economic losses. This study examines the application of artificial intelligence techniques for plant diseases. The proposed methodology includes several key steps. First, data augmentation techniques were employed to augment the available dataset of disease images, thereby enabling the model to learn more robust features. The augmented dataset is used to fine-tune four pre-trained deep learning models, namely VGG19, ResNet50V2, ResNet101V2, and ResNet152V2, which have been demonstrated to exhibit superior performance in image recognition tasks. A transfer learning approach was used, where pre-trained models were adapted to the plant disease classification task. This approach enabled the model to exploit knowledge from large-scale datasets while optimizing its performance on the target plant disease dataset. Finally, an extensive hyperparameter tuning process was conducted to identify the optimal configuration of the AI model, including parameters such as learning rate, units, optimizer, and dropout rate. This systematic approach to hyperparameter optimization allowed the model to achieve the best possible performance regarding plant disease accuracy. The results of this study demonstrate the effectiveness of the AI-based approach to automate plant disease detection, with significant implications for precision agriculture and sustainable food production.