The objective of this study was to establish and characterize the relationship between the dose of cloprostenol (37.5, 250, 500 and 750 μg) and the age of the early corpus luteum (CL) (80, 88, 96, 104 and 112 h) on the luteolytic response of mares. Behavioural oestrus and ultrasonographic signs of return to oestrus were considered as the occurrence of full luteolysis. A total of 298 mares were divided into groups according to dose of cloprostenol and CL age. There was an effect of dose of cloprostenol (p < 0.001) and age of the CL at the time of treatment (p < 0.001) on the percentage of mares with full luteolysis. The efficacy of 37.5 μg of d-cloprostenol was similar to that of 250 μg of d,l-cloprostenol (p > 0.05); and that of 500 similar to that of 750 μg (p > 0.05). The higher dose groups (500 and 750 μg) induced full luteolysis more frequently than the lower dose groups (37.5 and 250 μg) 96–104 h post-ovulation. There was no effect of CL age or cloprostenol dose on the interovulatory interval (p > 0.05). In conclusion, the effect of cloprostenol on the percentage of mares undergoing full luteolysis is dose-dependent. However, this effect is only evident in mares with CLs aged between 96 and 104 h. There is no advantage of administering more than 500 μg of d,l-cloprostenol (Estrumate®), to obtain a higher percentage of mares with full luteolysis in mares with CLs aged 80–112 h.