Chronic Kidney Disease (CKD) is an increasingly prevalent condition that poses significant challenges to global healthcare systems. The growing availability of medical data has opened new avenues for applying data mining techniques to support clinical decision making. This study investigates the use of classification methods, specifically Naive Bayes and Artificial Neural Network (ANN), for predicting CKD based on patient attributes. The dataset was analyzed using the RapidMiner platform, with preprocessing applied to ensure data quality and consistency. Comparative results indicate that the Naive Bayes classifier delivers higher accuracy and faster processing than ANN, making it a more suitable option for early-stage CKD detection. These findings demonstrate the value of lightweight and interpretable models for developing clinical support tools, particularly in low-resource environments.