Abstract: With the rapid development of intelligent classification techniques which depends on machine learning, this study addressed the comparison between one of the traditional statistical models (logistic regression) with the supervised machine learning model (support vector machine) in order to classify chronic kidney disease patients based on a blood test (serum) for a group of presence and absence patients. The dataset contains data of 153 cases and 11 attributes for diagnosis of chronic kidney disease. The dataset were divided into two groups (training and testing) and after applied the above models depend on evaluation performance criteria (model accuracy, model sensitivity, model specificity, prevalence, kappa coefficient and area under curve (ROC)). The study concluded the results indicate SVM Model is the best performer (best classifier). As well the study concluded through the final fitted models used that the most important factors that have a clear impact on chronic kidney disease patients are creatinine and urea.