Diabetes is a leading cause of blindness, renal failure, amputations, heart failure, and stroke, among other complications. When we eat, our bodies convert the food we eat into sugar, or gulcose. Diabetes mellitus is a type of chronic disease which increases the level of blood sugar. Diabetes affects almost 1.3 billion people worldwide, and many more are at risk. Many complications may occur if diabetes are not treated or fail to identify. The rise in machine learning approaches provides a helping hand to solve this critical problem. This work is intended towards creating a predictive model for the prediction of type 2 diabetes. As a result, various machine learning techniques are utilised to diagnose diabetes early on. The experiments use the PIMA Indian Diabetes Database (PIDD), which was available from the UCI machine learning repository. Using the 10 Folds cross-Validation approach and the ensemble method, the accuracy is tested over successfully and erroneously classified examples. When compared to basic classification algorithms, the results demonstrate that the ensemble technique outperforms with the greatest accuracy of 77.60%.