Objective To explore the influence of modified super-long hormone replacement protocol on pregnancy outcome of previous embryo implantation failure patients who underwent frozen-thawed embryo transfer (FET). Methods A total of 669 women who underwent FET with a failed history of in vitro fertilization/intracytoplasmic sperm injection-embryo transfer (IVF/ICSI-ET) were enrolled in Changhai Hospital of Second Military Medical University from Jan. 2015 to Jan. 2017. Based on different endometrial preparation protocols for FET, the patients were assigned to receive either modified super-long hormone replacement protocol (Modified super-long group, n=184) or conventional hormone replacement protocol (Conventional group, n=485).The pregnancy outcomes were retrospectively analyzed in the two groups. Results Patients in the two groups had similar baseline characteristics, including age, body mass index, duration of infertility, number of transferred embryo, endometrial thickness on transformation day and transplantation day (P>0.05). The good-quality embryo transfer rate in the Modified super-long group was significantly lower than that in the Conventional group (50.9% vs 64.8%, P<0.01), while the blastocyst transfer rate was significantly higher (57.4% vs 18.3%, P<0.01); compared with the Conventional group, the embryo implantation rate and clinical pregnancy rate in the Modified super-long group were both significantly increased (29.8% vs 23.8%, 46.7% vs 35.9%, respectively; both P<0.05). The logistics regression multivariate analysis in clinical pregnancy rate showed that the blastocyst transfer rate and good-quality embryo transfer rate were the influence factors for clinical pregnancy rate. Further subgroup analysis of different stage of transferred embryos showed that, in the cleavage embryo transferr cycles, the good-quality embryo transfer rate in the Conventional group was significantly higher than that in the Modified super-long group (67.5% vs 58.0%, P<0.05), but the embryo implantation rate and clinical pregnancy rate showed no significant difference between the two groups (23.0% vs 16.7%, 34.9% vs 32.0%, respectively; both P>0.05); however, in the blastocyst transfer cycles, although the good-quality embryo transfer rate had no significant difference between the two groups (45.5% vs 52.7%, P>0.05), the embryo implantation rate and clinical pregnancy rate in the Modified super-long group were significantly higher than those in the Conventional group (39.6% vs 27.2% and 56.9% vs 40.2%, respectively; both P<0.05). Conclusion The modified super-long hormone replacement protocol can improve the embryo implantation rate and clinical pregnancy rate of the patients undergoing FET with a history of embryo implantation failure, and is worthy of clinical popularization.