The objective of this study was to evaluate the relationship between maternal nutrition knowledge and maternal socio-demographics including participation in the Special Supplemental Women, Infants and Children's (WIC) Program. A cross-sectional study of new mothers at two San Francisco hospitals was conducted using some of the American Academy of Pediatrics' guidelines in a structured questionnaire to assess maternal nutritional knowledge. Maternal nutritional attitudes towards product nutrient labels were also assessed in a questionnaire format. Logistic regression models were used to evaluate the odds of having high maternal nutrition knowledge and of infrequently reading nutrition labels. In multivariate logistic regression models, higher maternal nutrition knowledge (defined as answering all four nutrition questions correctly) was associated with higher income levels defined as > or =$25 000/year, odds ratio (OR) 10.03 95% confidence interval (CI) (1.51-66.74), and in linear models, higher nutritional knowledge was associated with having more children (P < 0.01), a higher income (P = 0.01) and not being a WIC participant (P < 0.01). Mothers with higher incomes were also more likely to read product nutritional labels OR 4.24, 95% CI (1.24-14.51), compared with mothers with lower incomes as were mothers with higher education levels OR 3.32, 95% CI (1.28-8.63). In San Francisco, lower income mothers are at greatest risk for low maternal nutrition knowledge and not reading product nutritional labels. Higher household income was independently associated with increased maternal nutrition knowledge and likelihood of reading nutritional labels. More comprehensive interventions need to target low-income mothers including current WIC participants to help close the nutritional advantages gap conferred by income and education.