The objective of this study was to evaluate the effect of accessory corpus luteum (CL) induction on fertility in dairy cows. On day 5 after artificial insemination (AI), lactating Holstein cows were assigned unequally to receive gonadotrophin-releasing hormone treatment (GnRH) (n = 641) or no treatment (control; n = 289). Cows had their blood sampled for progesterone (P4), and ovaries were scanned by ultrasound on days 5, 12, 19, 26, 33, 47, and 61 after AI. Pregnancy diagnosis was performed on days 26, 33, 47, and 61. On day 12, cows treated with GnRH were allocated to ipsilateral (n = 239) or contralateral (n = 241) groups based on the side of accessory CL formation relative to previous ovulation. Accessory CL cows had greater P4 than controls. In total, 52.7% (78/148) of pregnant cows in contralateral group had accessory CL regression earlier (<day 33; 30.8%) or later (days 33–61; 69.2%) in pregnancy with coincident decrease in P4. No cows with ipsilateral accessory CL underwent regression. There was no difference in pregnancy/AI among groups. Cows with contralateral accessory CL that underwent early regression had greater pregnancy loss (30%) than controls (10%), or cows with ipsilateral CL (3%) or contralateral CL with either later or no regression (12%). Cows with ipsilateral accessory CL had lower pregnancy loss than controls. In conclusion, elevating circulating P4 by the induction of accessory CL, particularly ipsilateral CL, increases P4 and reduces pregnancy loss. However, contralateral accessory CL that undergoes regression before day 33 of pregnancy has increased pregnancy loss, possibly due to an abrupt decrease in P4 at a pivotal period of pregnancy (days 26–33).