Diabetes is a global health concern that affects the population around the world. In this study, the diabetes dataset is assessed using two machine learning approaches (Random Forest (RF) and Logistic Regression (LR) models). These approaches include the following frameworks: data collection process, preprocessing steps, and subsequent training and testing phases. The data set used here contains 768 instances with nine distinct features (pregnancy, glucose, blood pressure, BMI, insulin, skin thickness, diabetes pedigree function, age, and outcome). The parameter optimization of both models is done through fine-tuning. The primary evaluation focus on classification performance. Comparative analysis is done on both models capabilities in determining diabetic status. The findings indicated that the LR model achieved 78% accuracy while, the RF model achieved 72% accuracy. Overall the LR model performed marginally well in comparision with the RF model. This research helps to identify risk factors within a diabetes-related dataset, allowing effective classification of individuals into diabetic and non-diabetic categories.