The present study compared 2 strategies to initiate a progesterone (P4)-based timed artificial insemination (TAI) protocol for lactating dairy cows: only GnRH or estradiol benzoate (EB) plus GnRH (EB+GnRH). Lactating Holstein cows (n = 487; 184 primiparous and 303 multiparous) from 2 commercial dairy herds were used for their second or greater services postpartum. Each week, cows that were nonpregnant at the pregnancy diagnosis 32 d after a previous AI were randomly assigned to 1 of 2 experimental groups that differed only in the strategy to initiate (d 0) the TAI protocol. On d 0, every cow received a 2.0-g P4 implant; in the EB+GnRH group, cows were treated with 2.0 mg i.m. of EB and 16.8 µg i.m. of the GnRH analog buserelin acetate, whereas in the GnRH group, cows received only 16.8 µg i.m. of GnRH. On d 7 after the initial treatment, 0.530 mg i.m. of cloprostenol sodium (PGF) was administered in all cows, followed by a second dose on d 8, concomitant with 1.0 mg i.m. of estradiol cypionate and P4 implant withdrawal. The TAI was performed on d 10 (48 h after P4 device withdrawal) in both experimental groups. Only conventional Holstein semen was used throughout the study. The percentage of cows with corpus luteum (CL) on d 0 (73%) and overall ovulation rate after d 0 (54%) did not differ between groups. The CL regression between d 0 and the first PGF treatment was greater in the EB+GnRH group than the GnRH group (42% vs. 31%). Consequently, the proportion of cows with CL at PGF was greater when only GnRH was used on d 0 compared with EB+GnRH (86% vs. 82%), and the mean number of CL at PGF was greater (1.23 vs. 1.11). The expression of estrus near TAI was greater in GnRH group (84% vs. 77%), and cows showing estrus had greater (44% vs. 10%) pregnancy per AI (P/AI) on d 32 for both treatments. We found no effect of the presence of CL on d 0 or at PGF, nor of ovulation after d 0 or CL regression between d 0 and d 7 on fertility. However, fertility was critically impaired when cows did not have CL at both times, d 0 and at PGF treatment. We did not observe any interaction between treatment and other variables, and the P/AI was similar in cows receiving EB+GnRH or only GnRH on d 0 (37.8% vs. 36.6%). In summary, although there was no detectable difference in P/AI between treatments, this study demonstrated potential negative physiological outcomes caused by EB treatment on d 0 (greater incidence of luteolysis after d 0 and fewer cows with CL at PGF treatment). Overall, we found no benefit of adding EB at the initiation of a P4-based TAI protocol on fertility compared with using GnRH alone, despite differences in ovarian dynamics and expression of estrus.