This paper explores the possibility of using machine learning algorithms to predict type 2 diabetes. We selected two commonly used classification models: random forest and logistic regression, modeled patients’ clinical and lifestyle data, and compared their prediction performance. We found that the random forest model achieved the highest accuracy, demonstrated excellent classification results on the test set, and better distinguished between diabetic and non-diabetic patients by the confusion matrix and other evaluation metrics. The support vector machine and logistic regression perform slightly less well but achieve a high level of accuracy. The experimental results validate the effectiveness of the three machine learning algorithms, especially random forest, in the diabetes prediction task and provide useful practical experience for the intelligent prevention and control of chronic diseases. This study promotes the innovation of the diabetes prediction and management model, which is expected to alleviate the pressure on medical resources, reduce the burden of social health care, and improve the prognosis and quality of life of patients. In the future, we can consider expanding the data scale, exploring other machine learning algorithms, and integrating multimodal data to further realize the potential of artificial intelligence (AI) in the field of diabetes.