Diabetes is a chronic disease marked by elevated blood sugar levels that can seriously harm the kidneys, heart, eyes, brain, and renal system, among other organs. A range of data analysis methods are used in healthcare analytics to enhance patient care. In order to predict the illness as soon as possible and avoid it, a machine learning model was constructed using a variety of machine learning techniques. The research effort focuses on the 770 diabetic people living in the Andaman and Nicobar Islands. In order to preprocess the data for this study, the gathered dataset was divided into training and testing sets using exploratory data analysis approaches. The research then used feature engineering techniques to determine the importance of each characteristic and produce a precise diabetes mellitus prediction based on the risk factor determined by the Indian Diabetes Risk Score (IDRS). The model used in the current research work uses nine machine learning algorithms to produce accuracy, precision, recall, and F1 score: the Gaussian Naïve Bayes algorithms, Ada Boosting classifier, XG Boosting classifier, Random Forest classifier, Bagging classifier, Logistic Regression, Linear SVC Algorithm, KNN classifier, and decision trees algorithm. Based on the results, it was determined that the Gaussian Naïve Bayes technique, Decision Trees, Logistic Regression, Ada Boosting, XG Boosting, KNN, and Random Forest classifiers gave the best accuracy of 88%. The study then developed a hybrid ensemble gradient boosting technique and used it on the suggested system, producing the best outcomes in terms of ROC curve (87%), accuracy (99%), precision (79%), recall (84%), and AUC (87%). Thus, it can be concluded that the suggested hybrid ensemble gradient boosting classifier not only outperforms the other machine learning techniques and yields better results, but it also accurately predicts patients’ chance of acquiring diabetes mellitus based on risk factors.