Introduction: Preterm birth complication is the first leading cause of neonatal mortality in Ethiopia. In this study, the pooled prevalence and predictors of preterm mortality in Ethiopia were assessed. Methods: The research protocol was registered to PROSPERO under the CRD42023415334 registration number. PubMed, Google Scholar, Cochrane Library, and Hinari databases were searched for studies up to April 10, 2023. Cohort studies on the incidence of preterm mortality and its predictors in Ethiopia were included in this review. This review was reported using the preferred reporting items for systematic reviews and meta-analyses checklist. We extracted data from the selected papers and exported to R 4.2.3 and STATA version 15.0 for estimating pooled incidence proportion, density rate, and adjusted hazard ratio (AHR). By looking at the funnel plot and using Egger’s test, the publication bias was assessed. This study didn’t receive funds from any organization. Results: From 152 identified studies, 16 studies were selected for final analysis. The pooled incidence proportion and incidence density of preterm mortality were 30% (95% CI: 27%–32%) and 40 (95% CI: 30–40) per 1,000-person days, respectively. Being a male (AHR = 1.43, 95% CI: 1.09, 1.78), gestational age (0.83, 95% CI: 0.80, 0.87), born before 32 weeks (AHR = 2.15, 95% CI: 1.15, 3.84), antepartum hemorrhage (AHR = 2.46, 95% CI: 1.07, 3.84), the fifth minute Apgar rating less than seven (AHR = 1.84, 95% CI: 1.43, 2.25), birth asphyxia (AHR = 1.87, 95% CI; 1.59, 2.14), birth weight 1,000–1,499 g (AHR = 3.63, 95% CI: 1.69, 5.57), respiratory distress syndrome (AHR = 1.80, 95% CI: 1.59, 2.02), jaundice (AHR = 2.39, 95% CI: 1.16, 3.61), neonatal sepsis (AHR = 1.54, 95% CI: 1.25, 1.83), and not using kangaroo mother care (AHR =1.92, 95% CI: 1.11, 2.72) were predictors of time to death of preterm. Conclusions: Almost one third of preterm neonates treated at different health facilities in Ethiopia failed to survive to the first four weeks of life. Because every study that was included was an institution-based cohort study, the results may not generalized to preterm newborns who were not hospitalized at medical facilities.