Objective: To study the correlation of plasma vitamin A (VitA) levels between neonates and pregnant women in third trimester.Methods: A total of 688 pregnant women were recruited in Yuanshi and Laoting counties of Hebei Province, from May to June 2009. Venous blood samples of women before delivery and cord blood samples of newborns were collected and measured for retinol (retinol concentration was used to reflect VitA level) using high performance liquid chromatography assay. According to venous blood plasma retinol concentration, maternal VitA nutritional status was divided into deficiency (<0.70 μmol/L), marginal deficiency (0.70-<1.05 μmol/L), and sufficiency (≥1.05 μmol/L). According to cord blood plasma retinol concentration, neonatal VitA nutritional status was divided into deficiency (<0.35 μmol/L), marginal deficiency (0.35-<0.70 μmol/L), and sufficiency (≥0.70 μmol/L); neonatal VitA relative deficiency was further defined as cord blood plasma retinol concentration lower than the 10th percentile. VitA placental transport ratio was defined as retinol concentration in the neonates divided by that in pregnant women. Multivariable fractional polynomials (MFP) model and Pearson correlation were used to study the dose-response relationship between maternal and neonatal plasma VitA levels, Logistic regression model to estimate the effect of maternal VitA nutritional status on neonatal VitA deficiency, and MFP model and Spearman correlation to describe the relationship between maternal VitA level and VitA placental transport ratio.Results: The average retinol concentration of the pregnant women was (1.15±0.30) μmol/L, and the prevalence of VitA deficiency and marginal deficiency were 4.5% and 37.8%, respectively. Average retinol concentration of the neonates was (0.78±0.13) μmol/L, and no neonates were VitA deficiency, 28.2% of the neonates were marginal deficiency. After multivariable adjustment, the VitA level of the neonates was positively and linearly related to maternal VitA level (p^m=1, P<0.05), with the corresponding Pearson correlation coefficient of 0.13 (P<0.01). As compared with the women with sufficient VitA, those with VitA deficiency (crude OR=2.20, 95%CI:1.04-4.66) and marginal deficiency (crude OR=1.43, 95%CI:1.01-2.02) had higher risks to deliver neonates with VitA marginal deficiency; while the risks turned to be non-significant after multivariable adjustment. The pregnant women with VitA deficiency had higher risk to deliver neonates with relative VitA deficiency before and after multivariable adjustment (crude OR=3.02, 95%CI:1.21-7.50; adjusted OR=2.76, 95%CI:1.05-7.22). The maternal VitA level was negatively and non-linearly correlated with placental transport ratio (p^m= -0.5, P<0.05), with corresponding adjusted Spearman correlation coefficient of -0.82 (P<0.001).Conclusion: There was a positive linear dose-response relationship between VitA levels of newborns and pregnant women in third trimester, indicating that neonatal VitA storing levels at birth was affected by maternal VitA nutritional status.