Objective:To establish and verify a simple clinical prediction model for obstructive sleep apnea syndrome(OSAS)in children.Methods:The clinical data of 95 children aged 2-12 years underwent polysomnography(PSG)were screened.The subjects with OAHI≤1 were included into non-OSAS group(n=22)and those with OAHI＞1 into OSAS group(n=73).Gender,age,body mass index(BMI),night pulse minimum oxygen saturation(SpO2),tonsil grading and adenoid grading of the 2 groups were compared and analyzed.Binary Logistic regression analysis was used to statistically analyze the data and establish a clinical prediction model for OSAS in children.Results:There was significant difference in age,BMI,SpO2,tonsil grading and adenoid grading between the 2 groups(P＜0.05),there was no significant gender difference(P＞0.05).The model equation was as follows:X=2.366-0.769(age-continuous variable)+0.248(BMI-continuous variable)-3.413(SPO2-continuous variable)+2.104(tonsil grade Ⅲ-Ⅳ"yes").The result of internally validated Hosmer-Lemeshow test was P=0.612(P＞0.05),AUC was 0.821(0.713-0.929,P＜0.01),sensi-tivity was 83.3％,specificity was 76.4％.The accuracy of external validation was 73.8％,the AUC was 0.805(0.664-0.943,P＜0.01),the sensitivity was 84.6％and the specificity was 75％.Conclusion:The predictive model may have good predictive efficacy for 2-12 years old children with OSAS,and may assist clinicians in diagnosing children with OSAS.