The early detection of chronic kidney disease (CKD) is a significant challenge due to its progressive development. It frequently leads to fatalities and has a profoundly negative impact on people's health. However, early diagnosis and treatment can mitigate the severe repercussions of advanced CKD stages. The primary goal of this study is to suggest a comprehensive framework for predicting CKD by utilizing machine learning classifiers on a standardized CKD dataset. Researchers implemented a comparative analysis to assess the effectiveness of classification methods prior to and following min–max scaling. Notably, logistic regression exhibited the highest accuracy at 97.50%, with a remarkable 54% improvement in the K-neighbors classifier. Moreover, the passive-aggressive classifier demonstrated superior results, attaining an impressive 99% accuracy rate, as evidenced by the effectiveness of the classification methods pre- and post-feature extraction. Additionally, this classifier demonstrated substantial enhancements in sensitivity 4.26%, specificity 2.00%, and overall accuracy 3.92%. Increasing the scope of the proposed methodology to include additional medical conditions, including diabetes, cancer, and heart disease, has the potential to significantly improve patient outcomes and improve predictive diagnostic skills.