Aim: The aim of the present study was to establish Thai-specific reference ranges of triple markers for fetal Down syndrome as a function of gestational age as well as weight correction models and to compare the false positive rates when using Thai-specific model relative to Caucasian-specific model. Material and Methods: A total of 993 normal Thai pregnant women were determined for mid-trimester serum levels of alpha-fetoprotein (AFP), free-beta human chorionic gonadotropin (hCG), and unconjugated estriol (uE3), using DefiaXpress system (Perkin Elmer, Waltham, MA, USA). Results: The models of Thai-specific medians for AFP, b-hCG, and uE3, as well as the models for weight correction were derived and the normal reference ranges were constructed. The best fitted equation for AFP, b-hCG and uE3 are as follows: Predicted median = 2.675 × 10 (0.153 × GA in week), r = 0.979; 10 (-0.717 + 57.487/GA in week), r = 0.991; and 10 (5.678-69.346/GA in week), r = 0.997, respectively. The models were properly applied to another group of 302 Thai women, signifying that they were reliable models. The weight-adjusted gestation-specific medians derived from Caucasian models were significantly higher than those based on Thai models and the false positive rate could be reduced from 10 to 7.1% when Thai models were applied. Conclusion: Thai reference ranges of triple screen markers as a function of gestational age as well as weight correction models have been established. The Caucasian reference range, even after weight correction, gives a positive rate that is much higher than that it should be, strongly suggesting the need for ethnicity-specific medians. © 2012 2012 Japan Society of Obstetrics and Gynecology.