Objectives : To evaluate whether the day of delivery for women with regular menstrual history was predicted best from the last menstrual period (LMP), crown rump length (CRL) and or biparietal diameter (BPD). Methods : All of 561 women had estimated the day of delivery by LMP, CRL in the first trimester (In case of 217 women, it was available) and BPD in the second trimester. The accuracy of each method in predicting the day of delivery was determined. Those who were delivered after the spontaneous onset of labor were included. Differences among these methods were evaluated with nonparametric tests. Results : The percentage of women who delivered within 3 days of the estimated day of delivery was 254(45.3%) and 216(38.5%) of the women with pregnancies by BPD and LMP, respectively. And within 7days of the estimated day were 408(72.7%), 390(69.5%) of the women from BPD and the LMP. In the women, the BPD estimate was significantly better predictor within the 7days of the day of delivery than LMP estimate (p=0.027). Compared to LMP estimate, CRL and BPD estimates seemed to be advanced the day of delivery about 2.6 days in CRL, and 0.9 days in BPD (p=0.004, p=0.034). But we could not find any advantage of the CRL measurement in first trimester than single BPD measurement in the second trimester for the predictor of the day of delivery. Conclusions : When the difference between the methods in predicting the day of delivery was less than 7 days, the BPD measurement was better than the last menstrual period.