Increased interest in determining areas in need of improved food access led the U.S Department of Agriculture (USDA) to define food desert census tracts; however, no nationwide studies have compared dietary patterns in food desert tracts to other tracts. Our objective was to examine dietary patterns in residents of food desert and non-food desert census tracts. We performed a cross-sectional analysis of 19,179 participants in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study enrolled January 2003-October 2007. We used participants' geocoded address with USDA Food Desert Locator to identify food deserts and multivariable-adjusted odds ratios (ORs) to calculate adherence to Southern, Plant-based, and Mediterranean dietary patterns. Odds of adherence to the Southern dietary pattern were higher among white high school graduates (OR=1.41; 95% CI: 1.20-1.67), white college graduates (OR=1.91; 95% CI: 1.55-2.35) and black college graduates (OR=1.38; 95% CI: 1.14-1.68) who reside in a food desert versus non-food desert. Odds of adherence to the Plant-based dietary pattern were 15% lower among non-southeastern residents (OR=0.85; 95% CI: 0.72-0.99), who reside in food desert versus non-food desert. No statistically significant differences were observed for the Mediterranean dietary pattern. Residents living in food deserts had lower adherence to healthy dietary pattern than residents not living in food deserts; the association may vary by race, education, and region.