Abstract This study aimed to evaluate the association between legume intake and blood pressure, as well as the mediating role of cardiometabolic risk factors in patients in secondary cardiovascular prevention. Socio-demographic, anthropometric, clinical and food intake data were collected from the baseline of the multicentre study Brazilian Cardioprotective Nutritional Program Trial – BALANCE (RCT: NCT01620398). The relationships between variables were explored through path analysis. In total, 2247 individuals with a median age of 63·0 (45–91) years, 58·8 % (n 1321) male and 96·5 % (n 2168) with diagnosis of hypertension were included. Negative associations were observed between histidine intake and systolic blood pressure (SBP) (standardised coefficient (SC) = −0·057; P = 0·012) and between legume intake and BMI (SC = −0·061; P = 0·006). BMI was positively associated with triglycerides–glucose (TyG) index (SC = 0·173; P < 0·001), SBP (SC = 0·144; P < 0·001) and diastolic blood pressure (DBP) (SC = 0·177; P < 0·001), and TyG index was positively associated with DBP (SC = 0·079; P = 0·001). A negative indirect effect was observed between the intake of legumes, SBP and DBP, mediated by BMI (SC = −0·009; P = 0·011; SC = −0·011; P = 0·010, respectively). In addition, an indirect negative effect was found between the intake of legumes and the DBP, mediated simultaneously by BMI and TyG index (SC = −0·001; P = 0·037). In conclusion, legume intake presented a negative indirect association with blood pressure, mediated by insulin resistance (TyG) and adiposity (BMI) in individuals of secondary care in cardiology.