OBJECTIVE: To examine whether mortgage discrimination, or redlining, is a risk factor for preterm birth among African American women in Chicago, and how it is related to racial residential segregation. METHODS: This was a retrospective cross-sectional study in Chicago, Illinois, 1989-1991. African American mothers (n = 33,586) in the Illinois Transgenerational Birth File were linked to the 1990 census and the 1990-1995 Home Mortgage Disclosure Act database. Logistic regression models assessed the relationship between redlining and preterm birth rates. RESULTS: Preterm birth rates were higher among African American women in redlined areas (18.5%) vs. non-redlined areas (17.1%). Unadjusted and adjusted odds ratio for preterm birth among African American women in redlined neighborhoods, compared to non-redlined neighborhoods, were 1.08 (95% CI 1.03-1.14) and 1.12 (1.04-1.20), respectively. By level of racial residential segregation, preterm birth rate was elevated (18.2%) in redlined, high-proportion African American areas compared to non-redlined high-proportion African American areas (16.7%), redlined low- (16.2%) and mid-proportion (16.1%) African American areas. CONCLUSIONS: Mortgage discrimination may be an important measure of institutional racism to be used in understanding racial disparities in preterm birth.