Chronic periodontitis is a common oral disease that is a leading cause of tooth loss. This study develops a risk prediction system for severe chronic periodontitis using a mixed effects logistic regression model trained on longitudinal data from the EGAT2 cohort in Thailand. The model incorporates patient demographics, behaviors, dental features, and plaque score to predict risk. When validated, the model demonstrated high accuracy, sensitivity, and specificity over 90%, indicating it could help reduce the workload of dental examinations by identifying high-risk patients for targeted screening. However, further external validation and improved algorithms may enhance the model's performance and utility.