Plant diseases represent a significant threat to global agriculture, causing substantial crop losses and economic disruption worldwide. Traditional pest control methods often rely on harmful insecticides and pesticides, which negatively impact the environment and human health. Imaging and deep learning technologies offer a promising alternative for disease detection and classification. In this study, we proposed an optimized DenseNet architecture for enhanced plant disease detection and multiclass classification. To enhance detection accuracy, we preprocess images using Contrast-Limited Adaptive Histogram Equalization (CLAHE). To prevent overfitting and enhance model training, we utilized the Adam optimizer along with an early stopping mechanism. These techniques contributed to the model’s strong capability in accurately detecting plant diseases across different crop types. Experimental evaluations using PlantVillage dataset confirm the effectiveness of our approach, achieving promising performance (avg. accuracy of 99.81% and 99.78% for training-testing and validation set, respectively) for plant disease detection. By leveraging these advanced techniques, the model enables early disease detection, facilitating timely intervention and minimizing crop losses, ultimately supporting sustainable agricultural practices.