Objectives: The purpose of this study was to show a procedure for a random forest (RF) analysis which predicts periodontal disease status by using R and Orange Data Mining software, and helps us to understand how to apply the RF technique for dental research. Methods: Oral examination data of the 7th Korea National Health and Nutrition Examination Survey were used. A RF model was adopted to analyze the data where the target variable was periodontal disease status and the features were gender, age, education level, marital status, alcohol consumption level, smoking status, brushing before sleep, hypertension, and diabetes-related variables. Results: The important features of the RF analysis were in the order of age, marital status, and prevalence of hypertension and diabetes. The accuracy of the RF analysis was 73% which is not high enough for use in the clinical field. Conclusions: The RF technique is an ensemble method used to predict periodontal disease status which produces higher accurate outputs than a single method. This study provides a step-by-step guide using Orange Data Mining for researchers who want to study machine learning techniques.