Background and aimsAcute myocardial infarction (AMI) is one of the most prevalent illnesses endangering the elderly’s health. The predictive nutritional index (PNI) has been shown in several studies to be a good predictor of nutritional prognosis. In this study, we explored the correlation between PNI during hospitalization and the outcome of elderly AMI patients.MethodsElderly AMI patients in the Cardiac Intensive Care Unit of Huadong Hospital from September 2017 to April 2020 were recruited for analysis. The clinical and laboratory examination data of subjects were retrieved. All enrolled patients were monitored following discharge. The primary clinical endpoints encompass major adverse cardiovascular events (MACEs) and Composite endpoint (MACEs and all-cause mortality). Survival analyses were conducted via the Kaplan–Meier and the log-rank analyses, and the Cox, proportional hazards model, was employed for hazard rate (HR) calculation.Results307 subjects were recruited for analysis. The optimal PNI threshold is 40.923. Based on the Kaplan–Meier analysis, the elevated PNI group experienced better prognosis (P < 0.001). Cox analysis demonstrated that the PNI group was a stand-alone predictor for elderly AMI patient prognosis (HR = 1.674, 95% CI 1.076–2.604, P = 0.022). Subgroup analysis showed that the HR of the PNI group was the highest in the ST-segment elevation myocardial infarction (STEMI) subgroup (HR = 3.345, 95% CI 1.889–5.923, P = 0.05), but no discernible difference was observed in the non-ST-segment elevation myocardial infarction (NSTEMI) subgroup.ConclusionBased on our analyses, the PNI during hospitalization can accurately predict the prognosis of elderly STEMI patients but not that of elderly NSTEMI patients.