This study compares DenseNet-121 and Xception architectures for classifying Bell Pepper Bacterial Spot Diseases using deep learning techniques. To enhance disease detection accuracy, a novel investigation was conducted to evaluate the performance of these two widely used convolutional neural networks. Our analysis revealed that DenseNet-121 exhibited superior performance, achieving remarkable accuracy rates of 99 and 100% during the training, validation, and testing phases. The findings underscore the significance of selecting an appropriate deep-learning architecture for accurate disease classification, emphasizing the potential of DenseNet-121 in the context of plant pathology applications. This research contributes to the ongoing efforts to optimize disease detection models and supports advancing precision agriculture through state-of-the-art deep learning technologies.