Lack of complete regression of the corpus luteum (CL) after prostaglandin F_2α (PGF) treatment may reduce fertility during timed AI (TAI) protocols. A total of 373 lactating Holstein cows (172 primiparous; 201 multiparous) were synchronized with Double-Ovsynch starting at 53 ± 3 DIM (GnRH-7d-PGF-3d-GnRH-7d-GnRH-7d-PGF-56h-GnRH-16h-TAI). At the final PGF, cows were randomized to 1 of 2 treatments: 1PGF = No additional PGF treatment; 2PGF = Second PGF treatment 24h after first PGF. Blood samples were collected at the final PGF and GnRH treatments (72 and 16 h before timed AI). Only cows with P4 above 2.0 ng/ml before PGF (n = 344) were further analyzed. Fisher's exact test and t-test were used to analyze categorical and continuous variables respectively, and logistic regression analysis was used to calculate probabilities of CL regression and P/AI. Primiparous cows had greater P4 than multiparous cows (7.7 ± 0.22 vs. 6.5 ± 0.18; P = 0.0001) before treatment. At 56 h after PGF, P4 was greater (P = 0.005) for 1PGF (0.4 ± 0.04) than 2PGF (0.2 ± 0.05) cows. The percentage of cows with complete CL regression (<0.5 ng/mL at 56 h after PGF) was increased by second PGF for primparous (81.2% vs. 97.5%; 1PGF vs. 2 PGF; P = 0.001) or multiparous (84.4% vs, 96.7%; P = 0.006) cows. Cows with lower P4 at time of PGF had a reduced probability of complete CL regression with 1PGF by logistic regression (P = 0.02) or by comparing (P = 0.0016) quartile 1 (lowest P4; 2.0 to 4.8 ng/mL; 66.7%; 28/42) to the other 3 quartiles (88.1%; 118/134) which did not differ. In contrast, 2PGF cows had elevated CL regression (>95%) regardless of P4 at time of PGF (P = 0.60). Nonetheless, 2PGF increased CL regression for cows in both quartile 1 (66.7% vs. 95.1%; P = 0.0016) or quartiles 2–4 (88.1% vs. 97.6%; P = 0.0034). Interestingly, there was increasing P/AI with increasing P4 before PGF treatment for 2PGF (P = 0.02), but not 1PGF cows (P = 0.13). Cows with lower P4 at the time of PGF (Quartile 1) had similar P/AI for 1PGF (31.0%; 13/42) or 2PGF (31.7%; 13/41). In contrast, for quartiles 2–4 there was a tendency for decreased P/AI (P = 0.10) in 1PGF (44.0%; 59/134) vs. 2PGF (52.8%; 67/127) cows. Thus, treatment with a second PGF during Double-Ovsynch reduced inadequate CL regression and tended to increase P/AI, particularly in cows with greater P4 at the time of PGF treatment.