Ultrasound imaging is the initial prenatal screening test to analyze fetal development and to evaluate any abnormalities. In recent times, deep learning techniques have shown promising results in automating the classification of maternal-fetal planes from ultrasound images. However, the application of deep learning faces several challenges including the issue of class-imbalanced data. Our model uses a pre-trained Xception module fine-tuned for maternal-fetal plane classification. To address overfitting and data scarcity, we incorporate a data augmentation module. Recognizing the importance of interpretability in clinical settings, our work integrates explainable artificial intelligence techniques into the classification framework. By employing class activation maps, we provide clear visual explanations for the model’s decision-making process, enhancing trust and understanding among medical experts. With an accuracy of 99.81%, this model demonstrates exceptional performance while addressing key challenges in the field, potentially improving prenatal care through more reliable and interpretable ultrasound analysis.