BACKGROUND There is little research on whether normoresponsive patients who produced poor-quality embryos once verses those who produced poor-quality embryos twice when using a single COH protocol should change to a different controlled ovarian hyperstimulation (COH) protocol. MATERIAL AND METHODS In this retrospective study, we enrolled 108 patients with 1 PPOS failure who chose to continue receiving the progestin-primed ovarian stimulation (PPOS) protocol (n=61) versus those who decided to switch to the modified ultra-long protocol (n=47). We also enrolled 131 normoresponsive patients with 2 PPOS failures who chose to continue receiving the PPOS protocol (n=60) versus those who decided to switch to the modified ultra-long protocol (n=71) in the third cycle. RESULTS We found no significant difference in clinical outcomes of patients with 1 PPOS failure who continued using the PPOS protocol verses those who switched to the modified ultra-long protocol in the second cycle, expect for a lower cancelation rate (4.3% vs. 16.4%). However, the patients with 2 PPOS failures had significantly more good-quality embryos (0.9 vs. 0.4), more viable embryos (1.8 vs. 0.9), lower cancelation rates (18.3% vs. 53.3%), and higher pregnancy rates per aspirated cycle (26.8% vs. 10.0%) when switching to the modified ultra-long protocol compared to those who decided to continue receiving the PPOS protocol (P<0.05). Furthermore, the odds of clinical pregnancy (odds ratio [OR] 5.997, 95% confidence interval [CI] 1.476-24.361, P=0.01) were positively associated with switching to the COH protocol in the third cycle. CONCLUSIONS For normoresponsive patients with poor-quality embryos when using the PPOS protocol, switching to the modified ultra-long protocol after having 2 PPOS failures was associated with better ART outcomes.