The objectives of this study were to evaluate; 1) the stability of measured serum anti-Müllerian hormone (AMH) concentrations in samples after multiple freeze-thaw cycles, 2) the repeatability of AMH concentrations within mares during the same breeding season as well as across breeding seasons, and 3) the relationship between serum AMH concentrations and fertility (measured as first cycle pregnancy rates) in thoroughbred mares. For the first aim, AMH concentrations (n = 9) were examined across four freeze-thaw cycles with no significant change in measured AMH concentrations. For the second aim, serum AMH concentrations (n = 12) were examined over three successive estrous cycles and over two successive breeding seasons and AMH levels were significantly correlated for individual animals within (r; 0.71–82) or across breeding seasons (r = 0.81). For the third aim, Thoroughbred mares (n = 419) on farms in central Kentucky had blood samples taken during estrus. Pregnancy was determined with transrectal ultrasonography at Days 13–18 after mating and ovulation, and pregnancy outcome was recorded as open, pregnant or twins. The relationships between mare age, serum AMH concentrations and the interaction of age and AMH with pregnancy outcome was examined by nominal logistic regression, and the relationship between serum AMH concentrations and mare age, pregnancy outcome and the interaction of age and pregnancy outcome was examined by ANOVA. Data in this study were then stratified according to quartiles into lower (25%), mid-50% (second and third quartiles combined - 50%) and upper (25%) quartiles for age and serum AMH concentration for further analysis by logistic regression. There were significant effects of mare age and pregnancy outcome, but not their interaction on serum AMH concentrations which were higher (P = 0.04) in pregnant than in open mares (0.65 ± 0.03 vs 0.55 ± 0.04 ng/mL, respectively). Likewise, logistic regression revealed significant effect of mare age and AMH but not their interaction on pregnancy outcome on the first mated cycle. Mares in the lower AMH quartile were more likely to be open at Day 13–18 than mares in the middle (odds ratio (OR) = 1.87)=13 or upper quartile (OR = 2.62) for AMH concentrations. Mares in the mid-50% (OR = 3.91) or upper (OR = 4.97) age quartile were more likely to be open at Day 13–18 compared to mares in the young age quartile. Based upon a Chi-squared analysis, the proportion of pregnant mares differed across age quartiles (P < 0.0001) and was greater (P < 0.05) in the young mare quartile. The proportion of pregnant mares tended to differ across AMH quartile (P = 0.1), and when adjusted for age quartile using the Cochran-Mantel-Haenszel Test, the proportion of pregnant mares differed (P < 0.05) across AMH quartile. In conclusion, mares with peripheral AMH concentrations in the lowest quartile had lower fertility compared to mares in the mid-50% or upper quartile.