Accurate classification of jackfruit leaf growth stages and diseases is essential for effective crop management and early disease detection. This research proposes an advanced deep learning fusion framework that combines VGG16 and Vision Transformer (ViT), enhanced with fine-tuned feature extraction, to improve the classification accuracy of jackfruit leaf diseases and growth stages. The model integrates VGG16's convolutional layers for extracting spatial features and ViT's attention mechanism for capturing long-range dependencies, enabling the classification of visually similar symptoms. Experimental results demonstrate that the fusion framework achieves an overall accuracy of 98.50 %, outperforming individual models in terms of precision, recall, and F1-score. The fine-tuned feature extraction process further enhances the model's ability to accurately classify challenging categories like Pest_Damage and Senescence. These results confirm the framework's effectiveness, offering a promising solution for agricultural applications in disease identification. Future work will focus on optimizing the model for real-time deployment and extending its use to other crops and disease categories.