Independent association between legume intake and systemic inflammation is not well documented. The traditional Iranian diet provides an opportunity to assess the association between legume intake and health outcomes. This study was carried out to examine legume consumption in relation to serum concentrations of adhesion molecules and inflammatory biomarkers among Iranian women. In this cross-sectional study, 486 Tehrani female teachers were investigated. A trained dietitian administered a validated semiquantitative FFQ for assessment of usual dietary intakes. Legume intake was calculated by summing up the consumption of lentils, peas, chickpeas, different kinds of beans including broad beans, and chickling vetch. To measure serum concentrations of adhesion molecules and inflammatory biomarkers, a fasting blood sample was taken. After statistically controlling for potential confounders, individuals in the highest tertile of legume intake had lower serum concentrations of E-selectin (percent difference between the lowest and highest tertile: -14.1%; P = 0.04), soluble intercellular adhesion molecule-1 (-20.3%; P < 0.01), and soluble vascular cell adhesion molecule-1 (-15.6%; P = 0.01) compared with those in the lowest tertile. Legume intake was inversely associated with serum concentrations of high sensitive CRP (hs-CRP), TNFα, and IL-6, even after controlling for potential confounders and dietary variables (percent difference between the lowest and highest tertile for hs-CRP: -39.2%, P < 0.001; for TNFα: -15.9%, P = 0.04; and for IL-6: -39.5%, P < 0.01). Legume intake and concentrations of serum amyloid A were not correlated. Legume consumption is inversely associated with serum concentrations of adhesion molecules and inflammatory biomarkers among Iranian women.