Early disclosure of diabetes is crucial to evade veritable complications in patients. The reason of this work is to recognize and classify sort 2 diabetes in patients utilizing machine learning (ML) models, and to choose the primary idealize outline to anticipate the hazard of diabetes. In this paper, five ML models, counting K-nearest neighbor (K-NN), Bernoulli Naïve Bayes (BNB), choice tree (DT), calculated backslide (LR), and back vector machine (SVM), are reviewed to anticipate diabetic patients. A Kaggle-hosted Pima Indian dataset containing 768 patients with and without diabetes was utilized, checking components such as number of pregnancies the tireless has had, blood glucose concentration, diastolic blood weight, skinfold thickness, body assault levels, body mass record (BMI), intrinsic foundation, diabetes interior the family tree, age, and result (with/without diabetes). The comes around appear up that the K-NN and BNB models beat the other models. The K-N