Our objectives were to evaluate ovarian dynamics and fertility comparing 2 treatments at the start of a progesterone (P4)-based fixed-time artificial insemination (FTAI) protocol and 2 treatments at the end of the protocol. Thus, 1,035 lactating Holstein cows were assigned in a random phase of the estrous cycle to 1 of 4 treatments using a completely randomized design with a 2×2 factorial arrangement. At the beginning of the protocol (d -10), cows received GnRH or estradiol benzoate (EB) and, at the end, EB (d -1) or estradiol cypionate (ECP; d -2), resulting in 4 treatments: GnRH-EB, GnRH-ECP, EB-EB, and EB-ECP. All cows received an intravaginal P4 device on d -10, which was removed on d -2. Cows also received PGF2α on d -3 and -2. The FTAI was performed on d 0. Ovaries were evaluated by ultrasound for corpus luteum (CL) presence and regression (d -10 and -3) and follicle measurements (d -10 and 0), as well as the uterus for percentage pregnant per AI (P/AI; d 32 and 60). Blood samples were collected (d -10 and -3) for P4 measurements. Treatment with GnRH rather than EB tended to increase P/AI on d 32 (38.2 vs. 33.7%) and on d 60 (32.9 vs. 28.9%). More cows treated with GnRH had CL on d -3 compared with EB-treated cows (77.3 vs. 58.3%), due to less CL regression between d -10 and -3 (24.7 vs. 43.8%) and more cows with a new CL on d -3 (35.9 vs. 25.0%). Cows treated with GnRH also had greater P4 concentrations on d -3 than EB cows (3.4 vs. 2.0 ng/mL). Increased circulating P4 at the start of the protocol (d -10) decreased the probability of ovulation to EB or GnRH at that time. Cows from GnRH group also ovulated a larger-diameter follicle at the end of the protocol (15.5 vs. 14.7mm). No difference between EB and ECP in P/AI on d 32 (34.8 vs. 37.0) and 60 (30.8 vs. 31.0%) or in pregnancy loss (11.1 vs. 15.4%) was detected and we found no interaction between treatments for P/AI. Independent of treatment, cows with CL on d -10 and -3 had the greatest P/AI on d 60 (36.9%). In conclusion, treatments at the end of the protocol were similar for ECP or EB and we found no additive effect or interactions on P/AI between treatments. However, cows treated with GnRH rather than EB on d -10 had less luteolysis and tended to have greater P/AI, probably because P4 concentrations were greater during the protocol. Finally, regardless of treatments, cows with CL at the beginning of the protocol as well as at the time of PGF2α had greater fertility.