This study investigates the predictive power of serum progesterone/estradiol (P/E-2) level for estimating the live birth rate in patients who had a serum progesterone (P) rate >= 1.5 ng/mL on the human chorionic gonadotropin (hCG) administration day and who received the gonadotropin-releasing hormone (GnRH) antagonist protocol and intracytoplasmic sperm injection (ICSI). This retrospective cohort study included 176 cycles. The P/E-2 ratio was lower in patients with a live birth (0.73 +/- 0.54) than those without a live birth (1.05 +/- 1.38), but the difference was not statistically significant (p = .158). According to the receiver operating characteristic curve analysis of the hCG day P/E-2 ratio, the area under the curve was 0.579 (95% confidence interval: 0.478 - 0.680, p = .158) for predicting live birth. In conclusion, this study suggests that a P/E-2 ratio is not a significant predictor of a live birth rate in the patients with an hCG-day serum progesterone level of >= 1.5 ng/mL undergoing GnRH antagonist ICSI cycles with a fresh embryo transfer.