The aim of this retrospective cohort study was to investigate which preparation method is optimal for frozen-thawed embryo transfer (FET) treatment. Analyses were performed on 3160 FET cycles, including 654 cycles with a natural cycle (NC) protocol and 2506 cycles with an artificial cycle (AC) protocol. The primary outcome measures were the clinical pregnancy rate (CPR) and the live birth rate (LBR) per transfer. The Student's t-test, chi-square test and multiple logistic regression were used for statistical analysis. The CPR per transfer was 49.4% in the NC group and 58.6% in the AC group (OR = 1.270, 95% CI: 1.037-1.554). The LBR per transfer was 42.2% and 50.8% in the NC and AC groups, respectively (OR = 1.269, 95% CI: 1.037-1.552). Dividing the patients according to the type of transferred embryos, the CPR (67.3% versus 57.0%, p<0.01) and LBR (58.8% versus 49.7%, p<0.01) were higher after the AC protocol than after NC protocol in patients with blastocyst transfer. The NC and AC protocols yielded comparable CPR and LBR in the patients with cleavage embryo transfer. Our data indicate better pregnancy outcomes after the AC protocol than after the NC protocol. The AC protocol should be recommended in patients who were counseled before receiving FET treatment. Further studies are needed to confirm this finding.