Since the first detection of AIDS in Ethiopia, it claims about 67,000 lives of people in 2007. Ofrncourse, the introduction of ART dramatically improved survival and health quality of HIVinfectedrnpatients in the industrialized world; and the survival benefit of ART has been wellrnstudied too. However, in resource-poor settings, where such treatment was started only recently,rnlimited data exist on treatment results. Moreover, mortality has been high particularly in the firstrnmonth of initiating ART and factors contributing to this high mortality are poorly understood.rnThe study is designed to identify determinant factors (demographic and health related) that affectrnthe survival of HIV-infected patients in the Ethiopian Armed Forces based on the data fromrnArmed Forces General Teaching Hospital (AFGTH).rnA retrospective cohort study was conducted in AFGTH located in Addis Ababa, Ethiopia.rnRecords of patients enrolled between September 2003 and August 2007 were reviewedrncontinuously using patientsâ€™ ART unique identification numbers as reference. Kaplan-Meierrnsurvival curves and Log-Rank test were used to compare the survival experience of differentrncategory of patients, and proportional hazards Cox model was employed to identify independentrnpredictors of mortality. 734 patients on ART were followed for a median of 38.5 months (IQRrn10.75, 53). Of these 86 died during the follow up time of whom 28 (32.6%), 43 (50%) and 61rn(70.9%) deaths occurred within three months, six months and twelve months of ART initiation,rnrespectively. The independent predictors of mortality were low CD4 cell count at baseline, (HRrn= 0.995, 95% CI: 0.991 - 0.999), ambulatory and bedridden functional status, (HR=2.011,rn95%CI: 1.018 - 3.973) and (HR=3.358, 95%CI: 1.734 - 6.500), respectively, WHO clinicalrnstages III and IV (HR=7.052, 95%CI: 1.677- 29.658) and (HR=12.64, 95%CI: 3.003 - 53.199),rnrespectively, TB co-infection, (HR=1.734, 95% CI: 1.039 - 2.893) and OIs (HR=8.985, 95% CI:rn1.240 - 65.085).rnIn nut shell, there has been a high mortality of the cohort in the earlier months of treatment.rnThus, a careful monitoring of patients with low CD4 cell count, advanced WHO staging, historyrnof OIs, co-infection with TB and being employed is necessary in order to improve the survival ofrnAIDS patients