Periodontitis is a chronic inflammatory disease that affects millions of people worldwide which may result in severe consequences if not treated. Early detection and intervention can significantly reduce the risk of disease progression and related complications. Machine learning techniques have shown great promise in predicting periodontitis risk levels, offering a valuable tool for disease screening and management. This study aimed to evaluate the effectiveness of machine learning algorithms for predicting periodontitis risk levels, including multi-class and binary classification tasks. This study employed four distinct methods for handling missing values and utilized four different evaluation strategies, which included the hold-out validation, 5-fold cross-validation, leave-one-out cross-validation, and leave-one-patient-out cross-validation. In this study, we found using multinomial logistic regression with leave-one-patient-out cross-validation proved effective in predicting periodontitis risk levels. This algorithm exhibited high accuracy rates and consistent performance across all validation techniques with a train accuracy of 85% and test accuracy of 75%. Our findings provide valuable insights into the effectiveness of machine learning models for periodontitis prediction, highlighting the potential of these models for early detection.