Objective: To investigate the association between eating behaviour and metabolic risk in the broader population. Design: The association between metabolic risk factors (overweight, hypertension, hyperglycaemia, hypertriacylglycerolaemia, low HDL cholesterol, hyperuricaemia and fatty liver) and various eating behaviours were compared for four groups defined by subjective reporting: not eating until feeling full and not eating rapidly (G1); eating until feeling full only (G2); eating rapidly only (G3); and eating both rapidly and until feeling full (G4). Setting: A medical centre for health examinations in Tokyo, Japan. Subjects: Men (n 8240) and women (n 2955) who underwent health examinations. Results: The distribution of participants in G1 to G4 was 49?8 %, 11?5 %, 26?3 % and 12?4 % among men and 55?3 %, 15?0 %, 19?0 % and 10?7 % among women, respectively. Compared with G1, the age-adjusted OR (95 % CI) for overweight were significantly higher in G2 to G4, being respectively 1?85 (1?58, 2?17), 1?98 (1?76, 2?23) and 3?46 (2?99, 4?01) for men and 2?20 (1?62, 2?97), 2?59 (1?97, 3?39) and 3?12 (2?27, 4?26) for women. The age-adjusted OR were also significantly higher for hypertriacylglycerolaemia, hyperuricaemia and fatty liver in G2 and for all risks in G3 and G4 among men; and for hyperuricaemia in G2, for hyperglycaemia, hypertriacylglycerolaemia and fatty liver in G3 and for hypertriacylglycerolaemia and fatty liver in G4 among women. Conclusions: Both eating until feeling full and eating rapidly increase metabolic risk factors. Although the mechanism between rapid eating and metabolic risk requires further exploration, eating slowly and ending meals shortly before feeling full are important public health messages for reducing metabolic risk factors.