Disease in crop has a great impact on agriculture resulting in a loss. Early detection of diseases through regular observing is crucial to minimize crop loss. In recent years, automatic plant disease recognition systems have been developed using image. It represents several deep learning models, including VGG16, EfficientNetB0, ResNet-50, and GoogleNet, for the identification of four rice diseases utilization. The VGG16 model with a Dense Layer achieves an accuracy of 99.747%, while the VGG16 model with 2D-CNN achieves an accuracy of 99.916%. The EfficientNetB0 model with Dense Layer achieves a validation loss of 1.3880 and a test accuracy of 27.004%, while the EfficientNetB0 model with 2D-CNN achieves a validation loss of 1.3830 and a test accuracy of 27.004%. The ResNet-50 model with Dense Layer achieves a validation loss of 0.8121 and a test accuracy of 70.211%, while the ResNet-50 model with 2D-CNN achieves a validation loss of 0.7238 and a test accuracy of 69.789%. Finally, the GoogleNet model with Dense Layer achieves an accuracy of 99.916%, while the GoogleNet model with 2D-CNN achieves an accuracy of 99.831%. The above results are Extracted on PNG image dataset. The results show that the proposed deep learning models are effective in identifying rice diseases and provide a smart agriculture solution to the problem of crop diseases, helping farmers identify and manage the diseases efficiently, leading to better crop yield and quality. Overall, it demonstrates the impact of deep learning-based approaches for the automatic identification of plant diseases and the importance of using multiple models and datasets for comprehensive evaluation. These findings can aid in the development of more accurate and robust plant disease identification systems and help address the challenges of sustainable agriculture.