Blood sugar is harmful to people and their important organs, as it can cause blindness, renal disease, and heart and kidney problems. A lot of people worldwide are currently suffering from the harmful effects of diabetes with a large number of them going undetected at an early stage. Diabetes can lead to major health problems like blindness, cancer, stoke, knee pain, and kidney failure over time. Various researchers have attempted to develop an accurate diabetic prediction model. Researchers use different MLTs to address the issues and evaluate healthcare predictive analytics. Main focus of the research was how ML can be in use in diabetes. Based on these findings authors develop machine learning models and conceptual framework for diabetes prediction. Authors concentrated on detecting the diabetes on early stage. Researchers use framework to create and evaluate diabetes prediction models based on Radial Support Vector Machine (RSVM), Linear Support Vector Machine (LSVM), Naive Bayes (NB), Linear Regression (LR), Random Forest (RF), K-Nearest Neighbor (KNN), and Decision Tree (DT). Proposed work gives an accuracy of 0.64652 percent in RSVM. A K-fold cross validation (CV) approach was utilized to estimate seven different types of performance evaluation metrics.