Caffeine is consumed in various forms during pregnancy, has increased half-life during pregnancy and crosses the placental barrier. Small for gestational age (SGA) is an important perinatal outcome and has been associated with long term complications. We examined the association between maternal caffeine intake and SGA using National Birth Defects Prevention Study data. Non-malformed live born infants with an estimated date of delivery from 1997-2007 (n = 7,943) were included in this analysis. Maternal caffeine exposure was examined as total caffeine intake and individual caffeinated beverage type (coffee, tea, and soda); sex-, race/ethnic-, and parity-specific growth curves were constructed to estimate SGA births. Crude and adjusted odds ratios (aORs) and 95% confidence intervals were estimated using unconditional logistic regression. Interaction with caffeine exposures was assessed for maternal smoking, vasoconstrictor medication use, and folic acid. Six hundred forty-eight infants (8.2%) were found to be SGA in this analysis. Increasing aORs were observed for increasing intakes of total caffeine and for each caffeinated beverage with aORs (adjusting for maternal education, high blood pressure, and smoking) ranging from 1.3 to 2.1 for the highest intake categories (300+ mg/day total caffeine and 3+ servings/day for each beverage type). Little indication of additive interaction by maternal smoking, vasoconstrictor medication use, or folic acid intake was observed. We observed an increase in SGA births for mothers with higher caffeine intake, particularly for those consuming 300+ mg of caffeine per day. Increased aORs were also observed for tea intake but were more attenuated for coffee and soda intake.