Diabetic Retinopathy (DR) is a serious diabetic complication and the main cause of blindness globally. Preventing vision loss requires early DR lesion diagnosis and categorization. DR early detection and treatment can significantly reduce the risk vision loss .This paper focuses on classification of a sample into Diabetic & NonDiabetic, utilizing various techniques such as Decision Tree, ANN, KNN, SVM, Random Forest, Gradient Boosting Algorithms. The NCSU Diabetes dataset is pre-processed, and models are trained and evaluated for accuracy. SVM and ANN achieve over 80% accuracy, highlighting their potential in diabetes type classification. PIMA Indians Dataset is used for reference. The manual diagnosis process of DR retina fundus images by ophthalmologists is time-, effort-, and cost-consuming and prone to misdiagnosis unlike computeraided diagnosis systems. Recently, Machine learning has become one of the most common techniques that have achieved better performance in many areas like classification. To identify the best classifier for diabetic retinopathy SVM, Decision Tree, Logistic Regression, k-Nearest Neighbors, and ANN classifiers are compared in this paper. Furthermore, the DR available datasets have been reviewed. Various challenging issues that require more investigation are also discussed. Different machine learning algorithms are compared with the previous research and the results are satisfactory. This study improves diabetic retinopathy diagnosis by revealing the efficacy of different machine learning classifiers and helping develop accurate and efficient computer-aided diagnostic systems for early detection and management.