Objective: This prospective study assesses the efficacy of maternal serum screening for use in prenatal diagnosis of fetal anomaly and chromosome imbalance in Japanese women. Methods: Maternal serum α-fetoprotein, human chorionic gonadotropin, and unconjugated estriol were measured in 1,055 singleton pregnant women between 14 and 20 weeks of gestation. A calculated risk for trisomy 21 of ≥1/299 or α-fetoprotein ≥2.5 multiples of the median was adopted as positive. Results: Three hundred and seventy-eight of the 1, 055 women screened (35.8%) were identified as positive. Sensitivity, false-positive rate, and positive predictive value in women aged <35 years were 60.0, 10.6, and 6.8%, respectively, and these values were 87.5, 49.3, and 4.2%, respectively, in women aged ≥35 years. The false-positive rate in women aged <35 years was significantly lower than that for women aged ≥35 years (p < 0.001). Chromosomal abnormalities were identified in 21 cases, including 10 with trisomy 21, 5 with trisomy 18, 2 with trisomy 13, and 4 with other chromosomal disorders. Seventeen of the 21 cases (81.0%) showed screen-positive results, and among these all 10 cases with trisomy 21 were detectable. Two cases with trisomy 18, 1 with trisomy 13 and 1 with isochromosome X showed extremely low human chorionic gonadotropin levels (0.4 ± 0.1 multiples of the median, mean ± SE), although they were screen negative. Of the 264 women who did not undergo amniocentesis, none had any clinical findings consistent with aneuploidy after birth. Conclusions: Our results suggest that the evaluation of each serum marker, as well as of the calculated risk, was significantly important in the prenatal detection of fetal aneuploidy.