ABSTRACT Introduction: Acquired immune deficiency syndrome (AIDS) was the final stage of HIV infection. HIV attacks the CD4 cell which is responsible for the body's immune response to infectious agents. An infected person's CD4 cell count can be used to monitor the progression of the disease. The main objective of this study was to identify factors affecting the two outcomes namely CD4 cell count change and time to death for HIV/AIDS infected children under HAART at Felege Hiwot Comphrenesive Specialized Hospital. Methods: A retrospective cohort study design was conducted from 202 randomly selected HIV infected children in the ART clinic at FHRH under the follow-up period from January 2014 up to December 2018. The analysis consists of exploratory data analysis, the Kaplan-Meier survival estimate and Log-Rank test were used to compare the survival time and fitting three different models namely; a generalized linear mixed-effects model for the longitudinal data, a semi-parametric survival model for the time-to-event data and joint modeling of the two responses linked through their unobserved random effects. Results: The descriptive statistics revealed that about 25.2% of HIV infected children were died and 74.8% of HIV infected children were censored. The results for separate and joint models were quite similar to each other but not identical. However, the estimated association parameter ( ) in the joint model is -0.8339 and statistically significant (p-value =0.0254). This providing evidence that there was a very high association between the two sub-models. The relationship between the longitudinal CD4 count change and the risk of death was negatively associated. The result indicates that the higher value of CD4 count change was associated with a lower risk of death. Conclusions: The predictor education level, disclosure, functional status, and clinical stages were statistically significantly associated with the two responses of HIV/AIDS patients. When evaluating the overall performance of both the separate and joint models in terms of model parsimony, goodness of fit and the statistical significance of the association parameters, the joint model performs better. Thus, authors concluded that the joint model was preferred for the analysis of longitudinal measurement and survival data.