Gestational Diabetes Mellitus (GDM)is a serious condition that can affect pregnant women and the fetus. Early detection and management of GDM are crucial for preventing complications. In this paper, we propose a novel approach for early prediction of GDM using machine learning techniques, with a strong emphasis on explainability. Our methodology involves careful data preprocessing, feature selection, model evaluation, assessment and interpretability. By integrating explainable AI (XAI) principles, we ensure that our model’s predictions are transparent and trustworthy. Experimental results demonstrate the effectiveness of our approach, with the selected model achieving a high accuracy and F-score of 97.59%. Moreover, our method outperforms existing approaches, even with a reduced set of features, emphasizing the importance of feature selection in predictive modeling. The transparency provided by our explainability methods enhances trust and understanding of the model’s predictions. Our work lays the foundation for the development of a reliable and interpretable system for early detection of GDM, with implications for improving maternal and fetal health outcomes, and facilitating better decision-making by healthcare professionals.