In temperate climate zones, apple cultivation holds a pivotal position; however, foliar diseases pose substantial threats to orchard productivity and fruit quality. The current approach to disease diagnosis relies on labor-intensive manual examinations, necessitating a more efficient alternative. This study leverages the Plant Pathology 2021-FGVC8 dataset (R. Thapa, Q. Wang, N. Snavely, S. Belongie, Awais Khan, “The plant pathology 2021 challenge dataset to classify foliar disease of apples”, unpublished.), comprising approximately 19,000 high-quality RGB images of apple leaf diseases, to develop a deep learning model. Our primary emphasis centers on the utilization of the EfficientNet architecture for classifying diseases into multiple categories. Additionally, we conduct a comparative analysis, evaluating the performance of EfficientNet models to shed light on their efficacy in managing apple orchard diseases. The existing state-of-the-art research has explored different CNN models, such as R-CNN and CNN-Transformer-based models. Our work focuses on the application of EfficientNet on the plant pathology classification problem which seems promising since the training and inference can be performed with parameter counts that are orders of magnitude fewer than other models, and the EfficientNet models perform well in transfer learning scenarios. The inherent challenges posed by varying visual symptoms across apple cultivars and the influence of environmental factors during image acquisition underscore the significance of our research. This work represents a substantial stride toward automating the diagnosis of apple leaf diseases, holding the promise of transforming disease management practices within apple orchards.