Background. Evidence for the association between maternal thyroxine concentration and the risk of fetal congenital heart defects (CHDs) is absent. We aimed to study the association of maternal free and total thyroxine (FT4 and TT4) concentrations and the free-to-total thyroxine proportion (FTT4P, %) with the risk of CHD. Methods. The study was a hospital-based cohort study of 52,047 women who received a universal thyroid function test between 2012 and 2016. CHD was screened by ultrasound between 20 and 24 weeks of gestation or diagnosed until the 42nd day of birth. Adjusted odds ratios (ORs) of fetal CHD were estimated for maternal FT4 and TT4 concentrations or the FTT4P by multivariate logistic regression. Results. A total of 41,647 women with singleton pregnancies were included for the analysis and 215 CHD cases were detected. The FT4 concentration was significantly associated with a higher risk of CHDs (OR, 1.04, 95% confidence interval (CI): 1.01 to 1.07). Each 1% higher FTT4P was related to a 1.41-fold (95% CI: 0.27 to 3.59) higher risk of CHDs. The association became stronger for women with a thyroid function test performed between 12 and 18 weeks of gestation (OR = 1.05 (95% CI: 1.01 to 1.09) for the FT4 concentration and 3.32 (95% CI: 1.43 to 7.73) for the FTT4P). Conclusions. A higher FT4 concentration or FTT4P, measured between 12 and 18 weeks of gestation, was associated with an increased risk of CHDs. These findings may provide new insights into the mechanisms of CHDs and evidence for clinical decisions related to thyroid function tests.