The objectives were to evaluate changes in endometrial thickness (ET) near the time of a synchronized ovulation and to assess the relationship of ET and fertility in lactating Holstein cows, with or without estrogen supplementation near timed ovulation. In Experiment 1, eight cows were examined with transrectal ultrasonography, once daily for 5 d, starting concurrent with PGF 2α (PGF) treatment during an Ovsynch protocol (GnRH - 7d - PGF - 72h - GnRH). The ET increased rapidly after PGF (from ∼7 to ∼9.5 mm), remained > 9 mm for the next 2 d, then decreased to ∼8 and 7.4 mm, 1 and 2 d, respectively, after the second GnRH. In Experiment 2,642 cows (total of 758 breedings) were subjected to an Ovsynch protocol (GnRH - 7d - PGF - 56h - GnRH - 16h - timed AI); cows received either no further treatment (Ovsynch) or 1 mg of estradiol-17β im 8 h before the second GnRH (Ovsynch + E2). For both uterine horns, ET was measured (∼2 cm from the internal uterine body bifurcation) before E2 treatment (48 h after PGF). In cows with ET ≤ 8 mm vs > 8 mm, rates of ovulation were 86.0% (n = 136) vs 98.1% (n = 472; P < 0.01), respectively, and percentage pregnant per AI (P/AI) were 26.7% (n = 146) vs 42.7% (n = 524; P < 0.01). Treatment with E2 increased P/AI in cows with lower ET (Ovsynch + E2 = 37.0% vs Ovsynch = 23.3%; P = 0.07), but did not significantly improve P/AI in cows with ET > 8 mm (Ovsynch + E2 = 43.4% vs Ovsynch = 42.1%). In conclusion, a single ultrasonographic evaluation of ET in Holstein cows 48 h after PGF treatment in an Ovsynch program was a good predictor of ovulation failure and pregnancy success. Perhaps poor fertility in cows with reduced ET was low peripheral E2 concentrations near AI, poor P4 priming, or luteolysis failure during timed AI procedures.