Following kidney transplantation (KTx), renal function improves gradually until a baseline eGFR is achieved. Whether or not a recipient achieves the best-predicted eGFR after KTx may have important implications for immediate patient management, as well as for long-term graft survival. The aim of this cohort study was to calculate the renal function recovery (RFR) based on recipient and donor eGFR and to evaluate the association between RFR and long-term death-censored graft failure (DCGF). We studied 790 KTx recipients between January 1990 and August 2014. The last donor SCr prior to organ procurement was used to estimate donor GFR. Recipient eGFR was calculated using the average of the best three SCr values observed during the first 3 months post-KTx. RFR was defined as the ratio of recipient eGFR to half the donor eGFR. 53% of recipients had an RFR ≥1. There were 127 death-censored graft failures (16%). Recipients with an RFR ≥1 had less DCGF compared with those with an RFR <1 (HR 0.56; 95% CI 0.37–0.85; P = 0.006). Transplant era, acute rejection, ECD and DGF were also significant determinants of graft failure. Early recovery of predicted eGFR based on donor eGFR is associated with less DCGF after KTx.