Gestational diabetes (GD) is a form of diabetes that is first identified during pregnancy. It is a rapidly emerging condition among pregnant women and has become widespread in various populations around the world. Although its impact is considerable, there is currently no definitive cure; only symptomatic treatment is possible. This study aims to assess the risk of GD in women using modern data mining techniques for diagnostic purposes. The dataset was sourced from two well-known private hospitals in Dhaka, Bangladesh. We applied six classification algorithms: Support Vector Machine (SVM), Logistic Regression (LR), Decision Tree (DT), Gaussian Naive Bayes (GaussianNB), k-Nearest Neighbors (KNN), and Random Forest (RF) - both before and after implementing a feature selection step (selecting the top seven features), along with 5-fold cross-validation. In addition, a custom ensemble approach and two ensemble methods, bagging and boosting, were used. Our analysis revealed that the custom ensemble technique paired with the RF classifier achieved the highest accuracy of 88.24%. The findings of this research demonstrate strong potential to aid in the early detection and prevention of GD.