Diabetes, a global health challenge, affects millions worldwide, often leading to complications, such as organ damage, cardiovascular disease, and reduced quality of life. Among its forms, Gestational Diabetes Mellitus (GDM) emerges due to insulin resistance at the time of pregnancy. GDM may give rise to prolonged labor, cesarean delivery, and chronic health conditions such as increased chances of acquiring obesity and Type-2 diabetes and thus pose risks for both mother and child. In this study, GDM prediction has been investigated using advanced machine learning approaches on a comprehensive dataset. In this study, for predicting GDM, twelve machine learning algorithms have been implemented. Further meta-learning has been used with bagging approach to improve the performance of the model which has also improved the test data. Along with that, explainable AI techniques like SHAP and LIME analysis have been introduced to find the risk factors that aid in GDM prediction. The actual goal of this study is to build a comprehensive and reliable model for early prediction and diagnosis of GDM, thus bridging the gap between data science and healthcare. This approach might provide medical personnel with better resources for personalized treatment and better patient outcomes.