Little is known about the predictors recurrent ischemic events in patients with ST-segment elevation myocardial infarction (STEMI). This study aimed at investigating the predictors of recurrent myocardial infarction (MI) at long-term follow-up in a real-world STEMI cohort. All consecutive STEMI patients who underwent emergent coronary angiography and primary percutaneous coronary intervention between February 2013 and June 2019 at our institution were included. The primary outcome was recurrent MI; secondary outcomes were all-cause death, target vessel revascularization (TVR), in-stent restenosis, definite stent thrombosis (ST) and non-TVR. The study population included 724 STEMI patients; at median follow-up of 803 (324 to 1,394) days, the primary outcome was reported in 70 patients (10.1%). All-cause death occurred in 6.8%, TVR in 4.2%, in-stent restenosis in 2.5%, and ST in 1.9% of cases. At multivariable analysis, diabetes (hazard ratio [HR] = 1.18), serum level of lipoprotein(a) [Lp(a), HR = 1.01], and angiographic evidence of restenotic lesion (HR = 2.98) resulted independent predictors of recurrent MI. Kaplan-Meier analysis confirmed that diabetes, restenotic lesion, and differential Lp(a) risk range values, identified patients with lower long-term survival free from recurrent MI. Lp(a) level ≥ 30 mg/dL had an incremental prognostic stratification capability in patients with diabetes (HR = 5.34), and in patients with both diabetes and restenotic lesion (HR = 17.07). In conclusion, in this contemporary cohort of STEMI patients, diabetes, Lp(a) serum levels and restenotic lesions were independently associated with recurrent MI at long term. The coexistence of Lp(a) level ≥ 30 mg/dL showed an incremental risk stratification capability, supporting its implementation for long-term prognostic assessment in this high-risk clinical setting.