This study introduces a Deep learning framework of DenseNet201 for multiclass histopathological classification of bladder tissue into four classes: High-Grade Cancer (HGC), Low-Grade Cancer (LGC), Normal Stromal Tissue (NST), and Normal Transitional Layer (NTL). A total of 1,754 histopathological images were preprocessed by normalization, resizing, and noise reduction for consistency and high data quality. The dataset was divided into training, validation, and test data in a ratio of 80:10:10, and the DenseNet201 architecture – which was pre-trained on ImageNet – was finetuned using the Adam optimizer and binary cross-entropy loss for 20 epochs. The model had an overall accuracy of 89%, which showed good precision, recall, and F1-scores, especially in the HGC and NST classes. Confusion matrix analysis showed small misclassification between HGC and LGC categories, confirming that the discriminative power was high. These results make DenseNet201 a promising backbone for AI-assisted diagnostic systems with improved efficiency and interpretability for histopathological image analysis. Future work will explore data augmentation, attention mechanisms, and explainable AI to further improve model generalization and clinical applicability.