Abstract: Objective: To predict the early childhood caries（ECC） risk in Lingui District, Guilin by applying the preschool caries modifiedcaries assessment tool（PSC-MCAT）, and evaluate its predictive efficacy. Methods: A total of 367 3-year-old children from kindergartens in Lingui District, Guilin, were enrolled as subjects. Data were collected through baseline questionnaires and oral examinations, followed by a 1.5-year follow-up to observe the caries incidence. Chi-square trend tests were used to analyze the relationship between caries risk levels and caries incidence/dentin mean fracture time（dmft） increment. Generalized linear models were applied to assess new caries risk across different risk groups. Additionally, all PSC-MCAT items were used as predictive variables to construct caries prediction models using four machine learning algorithms—support vector machine（SVM）, logistic regression, naive Bayes, and random forest. Model performance evaluation validated the predictive value of the PSC-MCAT item combinations. Results: After 1.5 years of follow-up, statistically significant differences in new caries incidence and dmft increments were observed across caries risk groups（P＜0.05）. Chisquare trend tests revealed that caries incidence, dmft, and dmft increment significantly increased with higher caries risk levels（P＜0.001）. Generalized linear model analysis indicated that children in moderate-and high-risk groups faced significantly higher new caries risks than those in the low-risk group after 1.5 years（P＜0.05）. Among machine learning models based on the PSC-MCAT items, logistic regression and naive Bayes achieved AUC values ＞0.8 on the test set. All four algorithms（SVM, logistic regression, naive Bayes, and random forest） demonstrated accuracy, precision, sensitivity, and F1-value that exceeded 0.7. Conclusion: The PSC-MCAT demonstrates effective prediction of ECC risk. Machine learning validation confirms the reliability of its assessment items for ECC risk stratification, supporting its potential for clinical and public health implementation and application in Lingui District, Guilin.