Bangladesh, which ranks in the world’s top ten countries for production and consumption, depends significantly on rice for its economy and food needs. To ensure the healthy and proper growth of the rice plants, it is essential to detect any infections early on and before the affected plants receive the required treatment. An automated solution is inevitably prudent given the substantial time and labor costs involved in manual disease detection. This study offers an automated technique for correctly identifying eight types of rice leaf disease using a modified ResNet50-based transfer learning model, achieving an accuracy of 95.20%. Due to its superior accuracy and F1-score, which surpass those of newer models, ResNet50 was chosen. With its residual connections that boost feature learning, it is well suited for accurate agricultural image classification. When paired with drone and IoT technology, the system can provide real-time disease diagnosis, making it a more affordable option than manual detection.