The identification of apple leaf diseases is crucial to reduce yield reduction and timely take disease control measures. Employing machine learning-based methods, such as deep learning for accurate identification of multiple apple leaf diseases is challenging because of the limited availability of samples for supervised training and the serious class imbalance. Hence, this paper proposes an accurate deep learning-based pipeline to solve the problem of limited data sets on farms and reduce the partiality due to serious class imbalance. Firstly, this paper proposes the improved cycle-consistent adversarial networks (CycleGAN) to generate synthetic samples to improve the learning of data distribution and solve the problems of small data sets and class imbalance. On the basis of CycleGAN, two discriminators are introduced: one to judge whether the image is true or false, and one to judge whether the leaf image has disease. Specifically, healthy leaves are converted into disease-carrying leaves using two models, health-to-scab and health-to-rust to balance apple leaf disease datasets. Secondly, ResNet is trained as a baseline convolutional neural network classifier to classify apple leaf diseases. In experimental results, this paper carries out some experiments on evaluation of quality of the generated images by the improved CycleGAN and the performance of ResNet in terms of datasets and metrics. First of all, through qualitative observation of generated images and quantitative metrics, such as GAN-train and GAN-test, the generated diseases images from healthy image by the improved CycleGAN are superior to state of the art. Second, some experiments were carried out to verify the performance of the classification model, such as comparing five mainstream classification models and comparing them with traditional data augmentation methods. The results show that ResNet has the highest recognition accuracy on the test set, reaching 97.78%, and the classification accuracy is significantly improved by the generated synthetic samples (+14.7%). Finally, the experiment result of t-Distributed Stochastic Neighbor Embedding (t-SNE) visually confirmed that the images generated by improved CycleGAN have much better quality and are more convincing. Beyond that, a Visual Turing Test with three botanists showed that the generated images are indistinguishable from real apple leaf images.