This paper presents a novel approach to detecting plant leaf diseases by employing a deep stacking method that integrates the advantages of deep learning (DL) and traditional machine learning (ML) techniques. Convolutional Neural Networks (CNNs) serve as base learners, capturing intricate features from leaf images, while Gradient Boosting acts as a meta-learner to refine and enhance classification accuracy. The model completed training and assessment on a comprehensive dataset of plant leaf pictures, achieving impressive results getting an overall accuracy of 99.57% and significant improvements in validation loss across training epochs. The approach is validated through detailed evaluation metrics, such as precision, recall, and F1-scores, across multiple plant disease categories. The proposed method demonstrates high effectiveness in identifying and classifying various plant diseases, offering the potential for real-world agricultural applications in disease monitoring and prevention.