Objective: To explore the influencing factors of anxiety symptoms and to construct a predictive model based on machine learning algorithms, so as to provide support for the prevention and management of anxiety symptoms among junior high school students. Methods: From April to May 2023, a stratified random cluster sampling method was adopted to select 8 176 junior high school students from Zhengzhou and Shangqiu citys. All participants completed the Adolescent Self rating Life Events Checklist, the 10item Connor-Davidson Resilience Scale, the School Connectedness Scale, the Parent-Child Cohesion Questionnaire, and the 7 item Generalized Anxiety Disorder Scale. Logistic regression analysis identified the associated factors of anxiety symptoms among junior high school students. Predictive models were constructed using Logistic regression, Random Forest, and eXtreme Gradient Boosting (XGBoost) algorithms, with SHapley Additive exPlanations analysis explaining the optimal model. Results: The detection rate of anxiety symptoms among junior high school students was 16.3%. Logistic regression analysis showed that junior high school students who were female ( OR =1.22), in the ninth grade ( OR =1.27), living in urban areas ( OR =1.37), having a father with a college education or above ( OR =1.26), having a mother with a senior high school education ( OR =1.26), and experiencing higher levels of negative life events ( OR =1.05) reported a higher risk of anxiety symptoms(all P <0.05). In contrast, those with moderate family economic status ( OR =0.71), moderate academic burden ( OR =0.59), low academic burden ( OR =0.54), moderate sleep quality ( OR =0.46), good sleep quality ( OR =0.26), excellent sleep quality ( OR =0.15), higher levels of psychological resilience ( OR =0.96), higher levels of school connectedness ( OR =0.96), and higher levels of parent-child cohesion ( OR =0.98) reported a lower risk of anxiety symptoms (all P <0.05). Three machine learning models demonstrated good predictive performance for anxiety symptoms among junior high school students (all AUC>0.8), with the XGBoost model achieving the best predictive performance. SHAP analysis revealed that negative life events, sleep quality, school connectedness, psychological resilience and parent-child cohesion were the top five relevant factors for predicting anxiety symptoms. Conclusions The detection rate of anxiety symptoms among junior high school students is relatively high. The XGBoost model is the optimal predictive model for anxiety symptoms in the population. Negative life events, sleep quality, school connectedness, psychological resilience, and parent-child cohesion are significant correlates of anxiety symptoms among junior high school students.