Objective The aim of the current study was to evaluate the value of plasma endostatin for predicting 30-day mortality of patients with acute kidney injury (AKI). Methods Patients who underwent non-cardiac major surgery and developed AKI in the first 48 hours after admission to the intensive care unit were consecutively included. Concentrations of plasma neutrophil gelatinase-associated lipocalin (NGAL), cystatin C (Cys C), and endostatin were measured at three time points: 0, 24, and 48 hours after the AKI diagnosis. Clinical patient characteristics were recorded after AKI was diagnosed. Results A total of 256 new-onset AKI patients were enrolled. Of these, 48 (18.7%) patients died within 30 days. The difference in plasma endostatin values between 0 and 24 hours (ΔEndostatin-24h) yielded the best area under the curve (AUC) of 0.747 for predicting 30-day mortality in AKI patients; NGAL and Cys C achieved AUC of 0.672 and 0.647, respectively. The predictive AUC increased to 0.833 when ΔEndostatin-24h was combined with sequential organ failure assessment score and AKI classification. Conclusion Dynamic plasma endostatin is useful for predicting 30-day mortality in AKI patients. The predictive power of dynamic plasma endostatin can be significantly improved when it is combined with clinical patient data.