The objective of this study was to investigate the reliability of corpus luteum and dominant follicles, imaging parameters on the plasma progesterone, and prediction of pregnancy in cows. In total, 48 cows were used in this study and underwent Ovsynch program. Ultrasound imaging of the ovaries was done at the time of PG shots. Corpus luteum diameter (CL_d), area (CL_area), volume (CL_vol), and pixels (CL_PXL), as well as dominant follicle diameter (F_d) and area (F_area), were estimated using the ImageJ program. Blood samples were taken to assess progesterone (P4) concentrations. Pregnancy status was determined at 32 ± 3 days after insemination using an ultrasound “Sonoscape‐5 V.” Data were analysed using correlation analysis and ROC curves. Plasma P4 concentration showed positive correlation with CL_d (r = .68, p < .01), CL_area (r = .45, p < .01), CL_volume (r = .41, p < .01), and CL_pixels (r = .67, p < .01). The ROC curve indicated that P4 concentrations and CL parameters, especially the CL_pixels, were the best predictors of the pregnancy, among the others that were able to detect pregnancy at the time of PG with the best P4 and CL_pixels cut‐off value (4.1 ng/mL and 43.18) and AUC was (0.95 and 0.89); resp. (p < .001). Regarding the other parameters, it was possible to set AUC 0.79, 0.79, and 0.68 for CL_d, area, and volume with a sensitivity of 66.7%, 70.8%, and 66.7% and a specificity of 88.2%, 82.4%, and 70.4% (p < .01), respectively. The AUC for both the follicular diameter and F_area was (0.56) with a sensitivity of 58.33% and 70.8% and a specificity of 62.5% and 50% (p > .05), respectively. In conclusion, CL measurements and progesterone concentrations had the greatest pregnancy prediction in Holstein dairy cows.