The objectives of this study were (1) to investigate the effects of the preovulatory follicle (POF) size on the accuracy of Doppler-based early pregnancy detection, and (2) to determine whether the removal of PGF2α (PGF) treatment during the resynchronisation protocol would affect fertility in beef cows. In Experiment 1, Nelore suckling cows (n = 224) were enrolled in an estradiol-progesterone-based timed artificial insemination (TAI) protocol. At TAI, cows were separated based on the range of POF diameters, as follows: ≤11.0 mm (n = 50), 11.1-12.9 mm (n = 64), 13.0-14.4 mm (n = 62) and ≥14.5 mm (n = 48). On day 22 after TAI, the corpus luteum (CL) blood flow (CLBF) of all cows was examined by colour Doppler ultrasonography to diagnose nonpregnant cows. The cows with the largest POF had the greatest positive predictive value (88.6%; 31 of 35) and diagnostic accuracy (91.7%; 44 of 48). In Experiment 2, Nelore cows (n = 233) were subjected to the same TAI protocol. Fourteen days after TAI, all cows were started on a resynchronisation protocol. Cows diagnosed as nonpregnant based on CLBF, on day 22, received 0.5 mg estradiol cypionate intramuscular (im) and were assigned to receive either 150 μg of PGF (PGF; n = 50) or 2 mL of saline (control; n = 47). Cows treated with PGF had a P/AI of 30.0% compared with a 48.9% P/AI in controls (p = 0.06). Our findings demonstrate that the POF size affects the accuracy of a CLBF-based early pregnancy diagnosis and that the removal of PGF treatment from the resynchronisation protocol tended to increase P/AI of the second TAI.