---

Abstract
Introduction: High-risk behaviors predispose to delinquency in adolescents, and the identification of their predictions can be effective in preventive measures. The aim of this study was to determine the structural equation modelling of high-risk behaviors with loneliness and social network addiction mediated by academic burnout in adolescents.
Methods: The research method was descriptive-correlation. The second-high school students in Shiraz in 2020 were the statistical population of this study, 221 of whom participated in stratified random sampling method. Data collection instruments included demographic questionnaire, "Iranian Adolescents Risk-Taking Scale", "UCLA Loneliness Scale: Version 3", “Addiction to Mobile Questionnaire Based on Social Networks” and the “Maslach Burnout Inventory”. Construct validity was measured by confirmatory factor analysis and internal consistency was measured by Cronbach's alpha coefficient. Data were analyzed in SPSS. 24 and Amos. 24.
Results: The proposed model had a good fit. Also, the results of Bootstrap test showed that loneliness with mediation of academic burnout (β=0.085) and addiction to social networks with mediation of academic burnout (β=0.163) have an indirect and significant correlation with high-risk behaviors(P<0.05).
Conclusions: High-risk behaviors in adolescents based on loneliness, addiction to social networks can be predicted through mediation of academic burnout. It is suggested that psychologists and health professionals pay attention to the role of these variables.