Objective: To describe the experience of prenatal diagnosis for Hb Bart's disease, by chorionic villus sampling (CVS) with DNA analysis. Design: Descriptive study Settings: Department of Obstetrics and Gynecology, Faculty of Medicine, Chiang Mai University. Subjects: Sixteen high risk pregnancies at risk of Hb Bart's disease who were eligible for CVS criteria between 1 January, 1999 and May 31, 2000. Material and Method: Fetal villi were obtained by either transcervical (TC) or transabdominal (TA) CVS route to extract DNA and detect for α-thal-1 gene deletion (SEA type) with modified Chang's method. The CVS results were confirmed by either serial ultrasound or cordocentesis or diagnosis after pregnancy termination. Main outcome measures: The efficacy, safety and pregnancy outcomes. Results: CVS was successfully done in all of 16 cases (5 with TC and 11 with TA), The mean gestational age was 13.25 ± 2.9 weeks. The procedure time for TA was shorter than that of TC (4.64 ± 5.4 vs 10.4 ± 11.3 min). The CVS result showed as follows: 3 normal fetuses, 7 α-thal-1 carriers, 4 fetal Hb Bart's, 1 misdiagnosis and 1 failure to diagnosis due to technical error. The sensitivity and specificity were 100 per cent (4/4) and 90.91 per cent (10/11), respectively. One case of Hb Bart's misdiagnosis and one failure case were later confirmed for α-thal-1 trait and α-thal-1/Hb E trait by cordocentesis, respectively. The pregnancy outcomes included 11 livebirths, 4 terminated cases and 1 fetal loss of continuing pregnancies. No serious complications occurred. Conclusion: This preliminary experience suggests that CVS is an effective method for early prenatal diagnosis of fetal Hb Bart's.