BackgroundInternet Addiction is positively associated with a range of psychological risk factors such as childhood trauma and sleep disorders. However, it remains unclear if sleep duration mediates the association between childhood trauma and Internet addiction. MethodsWe enrolled 14,263 students from Shenzhen Polytechnic College, China. Sleep duration, Internet addiction and childhood maltreatment were assessed in these students by self-report measures, Internet Addiction Test (IAT) and Childhood Trauma Questionnaire (CTQ), respectively. With bootstrap approach and path analysis, the mediating role of sleep duration in the association between childhood trauma and Internet addiction was analysed. ResultsThe Internet-addicted group exhibited a higher level of the emotional abuse (EA) score, physical abuse (PA) score, sexual abuse (SA) score, a lower level of emotional neglect (EN) score and sleep duration compared with the control group (all p < 0.001). The CTQ total score and subscores showed a positive correlation with IAT scores both for males (r = 0.199, p < 0.001 for the total score, r = 0.356, p < 0.001 for EA, r = 0.270, p < 0.001 for PA, r = 0.249, p < 0.001 for SA, and r = 0.132, p < 0.001 for PN) and females (r = 0.127, p < 0.001 for the total score, r = 0.335, p < 0.001 for EA, r = 0.187, p < 0.001 for PA, r = 0.189, p < 0.001 for SA, and r = 0.065, p < 0.001 for PN). The CTQ subcores were negatively related to sleep duration both for males (r = -0.177, p < 0.001 for EA, r = -0.180, p < 0.001 for PA and r = 0.182, p < 0.001 for SA) and females (r = -0.137, p < 0.001 for EA, r = -0.105, p < 0.001 for PA, and r = -0.182, p < 0.001 for SA) and sleep duration was negatively correlated with IAT scores both in males (r = -0.120, p < 0.001) and females (r = -0.108, p < 0.001). Further, the path analysis suggested that EA and SA mediated significantly to the Internet addiction when all types of childhood trauma were examined in one model (both p < 0.001). ConclusionIn the current study, a great proportion of students met criteria for Internet addiction. Sleep duration mediated a significant proportion of the indirect effect between EA/SA and Internet addiction. The findings may help with prevention and intervention of Internet addiction in the future. The limitation of this study was that it was a cross-sectional study and not controlling for other mental disorders. Future large-scale longitudinal studies will be needed to further clarify the relationship between childhood abuse and Internet addiction and the mediation role of sleep duration.