This paper addresses the pressing issue of diabetes, a primary global health concern, by comparing machine learning models and a Centers for Disease Control and Prevention (CDC) questionnaire for diabetes diagnosis based on nutritional data. Utilizing the National Health and Nutrition Examination Survey (NHANES) dataset, the study highlights the limited of Machine Learning (ML) research focusing solely on nutritional features. The machine learning models, including K nearest neighbor (KNN) and support vector machine (SVM), are evaluated for recall, precision, and other performance metrics. Notably, KNN and SVM outperform the CDC questionnaire, obtaining F1 scores of more than 60% compared to about 40% for the questionnaire. In addition to that, they achieved recall scores of more than 80%, while the questionnaire only managed a recall of about 30%. These results suggest that machine learning models based on nutrition and diet could serve as more effective diabetes screening tools. © 2024 Copyright held by the owner/author(s).