The classification of medical images presents significant challenges due to the presence of noise, uncertainty, and indeterminate information. Traditional deep learning models often struggle to manage this, leading to reduced diagnostic accuracy, especially when dealing with low-quality or ambiguous conditions. This paper proposes a hybrid approach that integrates Neutrosophic Set (NS) theory with deep learning models to enhance X-ray image classification under uncertain conditions. NS theory introduces three domains: True (T), Indeterminate (I), and False (F) to manage image uncertainty and noise, allowing deep learning models to better interpret complex, ambiguous visual information. To evaluate the approach, five state-of-the-art deep learning models—MobileNet, ResNet50, VGG16, DenseNet121, and InceptionV3 are utilized, and their performance was evaluated on two different medical image datasets: Cervical spine injuries detection and chest disease classification. The results indicate that models trained on NS-transformed data, particularly DenseNet and MobileNet, yield superior outcomes compared to those trained on the original data, achieving significantly higher accuracy, precision, and recall. This demonstrates that incorporating NS theory into deep learning models significantly enhances their ability to classify uncertain and noisy X-ray images, providing a robust solution for improving diagnostic accuracy in medical imaging.