Maize is a crucial crop grown in India, and maize leaf diseases often result in significant yield losses. Therefore, it is crucial to identify these diseases and provide remedies for them. In this study, we propose using the Extreme Gradient Boosting (XGBoost) classifier along with extracted Convolutional Neural Network (CNN) features to detect and classify various maize diseases. To enhance the model’s generalization capability, we employ augmentation techniques. For training and testing purposes, publicly available datasets have been used. The CNN model alone achieves a test accuracy of 90.8%. However, by training the extracted CNN features with the XGBoost classifier, the accuracy further improves to 92.5%. Additionally, we introduce an Android-based solution that offers farmers a convenient method to identify crop diseases.