OBJECTIVE To study the relationship between dietary vitamin A intake and plasma vitamin A concentration, and establish the theoretical basis for dietary intake predicting vitamin A nutritional status. METHODS By using cluster sampling, 492 children aged 2-7 years in kindergartens in Banan district of Chongqing were selected. A cross-sectional nutrition and health survey was conducted, including the clinical examination, anthropometry, laboratory test and dietary survey. RESULTS Among the children surveyed, 229 were boys, and 263 girls, the mean age was (4.54 ± 0.87) years, height (107.50 ± 7.20) cm, and weight (18.42 ± 3.41) kg, the mean value of plasma vitamin A was (1.04 ± 0.30) μmol/L. The prevalence of marginal vitamin A deficiency (MVAD) was 43.5%. No cases of severe clinical vitamin A deficiency were found (plasma vitamin A ≤ 0.35 μmol/L). Clinical examination found no conjunctiva, corneaor skin abnormalities, and no Bitot's spots. Prevalence of the last two weeks colds were 27.4% (135/492), no diarrhea and other gastrointestinal or digestive diseases were found. The proportion of insufficient dietary vitamin A intake (<600 μg RE/d) was as high as 50.0%. By using correlation analysis, plasma retinol concentrations were related to dietary vitamin A intake (r=0.162, P<0.001), and to dietary energy intake (r=0.107, P=0.017). After adjustment for the effects of other non-dietary factors on vitamin A deficiency, the multivariate logistic regression showed that vitamin A-rich foods of liver intake=0 g/d (OR=1.95, 95% CI: 1.05-3.61, P=0.034), vitamin A-rich fruits intake=0 g/d (OR=1.55, 95% CI: 1.03-2.33, P=0.034), vitamin A-rich vegetables intake<200 g/d (OR=3.47, 95% CI: 1.37-8.75, P=0.009) were important risk factors of vitamin A deficiency. But we had not found the correlation between the intake of meat, eggs and milk and vitamin A deficiency. CONCLUSIONS Dietary factors may be the major risk factor of vitamin A deficiency in the three kindergartens. The dietary vitamin A intakes are significantly related to plasma retinol concentrations, and the vitamin A-rich foods intakes can predict the body's vitamin A nutritional status.