In recent years, plant diseases have caused significant damage to crop production, leading to huge economic losses. Therefore, the early detection of plant diseases is essential to minimize the damage and increase crop yield. In this paper, we present a comparative study of two popular deep learning models, VGG16 and ResNet50, for the detection of plant diseases. we compared the performance of these two models in detecting diseases in four crops, namely tomato, cotton, rice, and sugarcane. Our experimental results show that VGG16 achieved higher accuracy than ResNet50 for all four crops, with an average accuracy of 91.2% compared to 88.3% for ResNet50, demonstrating its effectiveness in plant disease detection.