The objective was to determine the effect of 2 presynchronization treatments on first-service pregnancy per artificial insemination (P/AI) in 4 dairy herds during warm and cool seasons of the year. Cows with ear tags ending with even digits at calving were enrolled in Presynch-10 (Presynch-10): two 25-mg injections of PGF2α (i.e., PG-1 and PG-2) 14 d apart. Cows with ear tags ending with odd digits were enrolled in PG-3-G: one 25-mg injection of PG (Pre-PG) 3 d before injection of 100μg of GnRH (Pre-GnRH), with the Pre-PG injection administered at the same time as PG-2 in the Presynch-10 treatment. Ten days after PG-2 or Pre-PG, all cows were enrolled in a timed AI protocol (Ovsynch-56; injection of GnRH 7 d before GnRH-1 and 56 h after GnRH-2 PG with AI 16 to 18 h after GnRH-2). Median days in milk (DIM) at scheduled timed AI were 75 d, which did not differ among herds. Cows detected in estrus before the scheduled timed AI were inseminated early (early bred, EB). Pregnancy was diagnosed at d 32 to 38 and at d 60 to 66 after timed AI by transrectal ultrasonography or transrectal palpation. Data were analyzed with herd as a random effect and with fixed effects of treatment (EB, Presynch-10, or PG-3-G), parity (primiparous vs. multiparous), season [hot (June through September) vs. cool-cold (October through May)], DIM, estrus at timed AI (0 vs. 1), and all 2-way interactions with treatment. The P/AI at d 32 to 38 for EB (n=472), Presynch-10 (n=1,247), and PG-3-G (n=1,286) were 31.4, 35.0, and 41.2%, respectively; P/AI at d 60 to 66 was 29.8, 32.2, and 37.3%, respectively. Season significantly influenced P/AI at d 32 to 38 and d 60 to 66, but a treatment by season interaction was not detected. The P/AI for PG-3-G and Presynch-10 treatments did not differ during cool-cold weather (d 32 to 38: 46.8 vs. 44.3%; d 60 to 66: 41.6 vs. 41.1%, respectively), but PG-3-G and Presynch-10 produced more P/AI than EB at d 32 to 38. During the summer, P/AI in PG-3-G was greater than in Presynch-10 (d 32 to 38: 35.9 vs. 26.7% and d 60 to 66: 33.2 vs. 24.4%, respectively), and P/AI in EB cows did not differ from that of Presynch-10 cows. Although pregnancy loss did not differ for EB, Presynch-10, and PG-3-G treatments (4.0, 6.7, and 9.3%, respectively), pregnancy loss from d 32 to 38 and d 60 to 66 was 2-fold greater in thinner cows (<2.5 vs. ≥2.5; 9.0 vs. 4.4%). We concluded that presynchronizing estrous cycles with PG-3-G produced more P/AI than inseminating cows at estrus during cooler weather and was superior to Presynch-10 during the summer.