Serum alpha‐fetoprotein (AFP) has been increasingly recognized as a marker for a poor prognosis after liver transplantation (LT) for hepatocellular carcinoma (HCC). Many published reports, however, have included a large proportion of patients with HCC beyond the Milan criteria, and the effects of incorporating AFP as an exclusion criterion for LT remain unclear. We studied 211 consecutive patients undergoing LT for HCC within the Milan criteria according to imaging under the Model for End‐Stage Liver Disease organ allocation system between June 2002 and January 2009. The majority (93.4%) had locoregional therapy before LT. The median follow‐up was 4.5 years (minimum = 2 years). The Kaplan‐Meier 1‐ and 5‐year patient survival rates were 94.3% and 83.4%, respectively. In a univariate analysis, significant predictors of HCC recurrence included vascular invasion hazard ratio (HR) = 10, 95% confidence interval (CI) = 3.9‐26, P < 0.001, a pathological tumor stage beyond the University of California San Francisco criteria (HR = 4.1, 95% CI = 1.36‐12.6, P = 0.01), an AFP level > 1000 ng/mL (HR = 4.5, 95% CI = 1.3‐15.3, P = 0.02), and an AFP level > 500 ng/mL (HR = 3.1, 95% CI = 1.04‐9.4, P = 0.04). In a multivariate analysis, vascular invasion was the only significant predictor of tumor recurrence (HR = 5.6, 95% CI = 1.9‐19, P = 0.02). An AFP level > 1000 ng/mL was the strongest pretransplant variable predicting vascular invasion (odds ratio = 6.8, 95% CI = 1.6‐19.1, P = 0.006). The 1‐ and 5‐year rates of survival without recurrence were 90% and 52.7%, respectively, for patients with an AFP level > 1000 ng/mL and 95% and 80.3%, respectively, for patients with an AFP level ≤ 1000 ng/mL (P = 0.026). Applying an AFP level > 1000 ng/mL as a cutoff would have resulted in the exclusion of 4.7% of the patients from LT and a 20% reduction in HCC recurrence. In conclusion, an AFP level > 1000 ng/mL may be a surrogate for vascular invasion and may be used to predict posttransplant HCC recurrence. Incorporating an AFP level > 1000 ng/mL as an exclusion criterion for LT within the Milan criteria may further improve posttransplant outcomes. Liver Transpl 20:945‐951, 2014. © 2014 AASLD.