The aim of the present study was to compare pregnancy and embryo loss rates in non-lactating mares bred either in the first, or in other estrus cycles during the breeding season. A total of 324 non-lactating mares, 179 of them barren and 145 maiden, were allocated in a stallion management center in Southern Brazil. All the mares were submitted to an artificial photoperiod of 15 hours light/9 hours darkness, either from May 15th on, or later on, from the moment they entered the management center. All the mares have been examined twice a week, from July 15th to July 31st, in order to determine the estrus phase, being classified, according to the findings, into two groups: group 1 = 1st estrus-mares showing reduced uterine tonus, small sized ovaries or ovaries with irregular follicular activity, lacking evidence of preovulatory follicle and, mainly, of corpus luteum (CL) at ultrasound; group 2 = ∆ 2 nd estrus-mares presenting CL at examination. The last ones received Prostaglandin on the 5th day after CL identification, or later on, for induction of a new estrus cycle. From August 1st on, all mares in estrus have been examined at maximal intervals of 48 hours, in order to evaluate follicular growth, grade of uterine edema and presence of intra-uterine fluid (IUF). Mares have been bred when a preovulatory follicle and low grade of uterine edema were found. A new examination was performed 36 to 48 hours after breeding to confirm ovulation and to detect IUF. Mares showing IUF were treated by means of uterine flushes and antibiotic infusion and those showing IUF for more than three consecutive days were excluded from the experiment. Barren and maiden mares bred in the 1st estrus cycle showed significant lower (p<0,026) pregnancy rates in days 12 and 42 in comparison to mares bred during subsequent cycles. Barren mares bred during the 1st estrus cycle showed higher embryo loss rate (p= 0.048) than barren mares bred during other cycles (20.4% vs. 8.0%). Embryo loss rate was not significantly different in maiden mares bred in 1st or in other estrus (9.1% vs. 4.9%). Grade of uterine edema and presence of IUF during estrus did not differ in barren and maiden mares bred either in the 1st, or in other cycles. However, barren mares bred during the 2nd and other cycles showed higher incidence of IUF (p=0.040) 36 to 48 hours after breeding than those bred in the 1st estrus cycle (51.7% vs.27.5%). Presence of IUF was not significantly different (p=0.680) between maiden mares bred during the 1st or other cycles (18,2% vs. 27.0%). It was concluded that: (a) during the breeding season the fertility of non-lactating mares is lower in the 1st than in other estrus cycles; (b) that a higher embryo loss rate may be expected in barren mares bred in the 1st than in other estrus cycles of the breeding season and (c) that the lower fertility rate observed during the 1st estrus cycle is not related to the incidence of post-breeding endometritis.