IntroductionHighly active antiretroviral therapy (HAART) was piloted in 2002 and was scaled up in 2003 in mainland China. The aim of this study was to evaluate the mortality and its possible predictors based on the long-term initial antiretroviral therapy (ART) cohort among HIV positive children and adolescents. MethodsThis prospective open-labeled multicenter cohort study was conducted from January 2008 to July 2021. The participants were recruited from six representative sites in mainland China. A total of 609 participants with an HIV-positive serostatus and <18 years old were recruited and each participant was informed consent at the time of enrollment. Mortality and annual hazard were calculated, and predictors for death were analyzed using Cox regression models generating hazard ratios (HR). ResultsThe results showed that the mortality was 0.721 per hundred person-years, and the annual hazard was less than 0.10 over time. Both CD4+T cell count and CD4+T cell percentage declined in the death group during the follow-up. The Cox regression model showed that the baseline low CD4+T cell count level (Low vs. High: aHR = 8.309, 95% CI: (1.093, 63.135)) and age >5 years old at HIV diagnosis (6–12 vs. 0–5: aHR = 3.140, 95%CI: (1.331, 27.411)); 13–18 vs. 0–5: aHR = 5.451, 95%CI: (1.434, 20.724)) were possible risk factors for death. ConclusionThe longitudinal cohort study demonstrated the efficacy of China's ART program among HIV-positive children and adolescents which could be beneficial to other countries with limited resources.