Background: Merging studies have reported the association of lipoprotein(a) [Lp(a)] with poor outcomes of coronary artery disease (CAD) in patients with type 2 diabetes mellitus (T2DM). However, the prognostic importance of Lp(a) for recurrent cardiovascular events (CVEs) is currently undetermined in patients with T2DM and prior CVEs. Methods: From April 2011 to March 2017, we consecutively recruited 2284 T2DM patients with prior CVEs. Patients were categorized into low, medium, and high groups by Lp(a) levels and followed up for recurrent CVEs, including nonfatal acute myocardial infarction, stroke, and cardiovascular mortality. Kaplan-Meier, Cox regression and C-statistic analyses were performed. Results: During 7613 patient-years' follow-up, 153 recurrent CVEs occurred. Lp(a) levels were significantly higher in patients with recurrent CVEs than counterparts (20.44 vs. 14.71 mg/dL, p = 0.002). Kaplan-Meier analysis revealed that the event-free survival rate was dramatically lower in high and medium Lp(a) groups than that in low group irrespective of HBA1c status (< 7.0%; ≥ 7.0%, both p < 0.05). Furthermore, multivariate Cox regression models indicated that Lp(a) was independently associated with high risk of recurrent CVEs [HR(95% CI): 2.049 (1.308-3.212)], such data remains in different HBA1c status (HR(95% CI): < 7.0%, 2.009 (1.051-3.840); ≥ 7.0%, 2.162 (1.148-4.073)). Moreover, the results of C-statistic were significantly improved by 0.029 when added Lp(a) to the Cox model. Conclusions: Our data, for the first time, confirmed that Lp(a) was an independent predictor for recurrent CVEs in T2DM patients with prior CVEs, suggesting that Lp(a) measurement may help to further risk stratification for T2DM patients after they suffered a first CVE. © 2020 The Author(s).