Estimating intake of phyto-oestrogens (PO) is difficult because there is inadequate information on the PO content of foods. Development of a biomarker of intake is therefore necessary for carrying out epidemiological studies. We aimed to validate a newly constructed PO database, containing more than 600 values assigned to foods by using duplicate diet analysis, and to investigate the relationships between measured PO intake, urinary excretion and plasma concentrations of PO. Fourteen subjects with estimated dietary intakes of PO ranging from 0 to 44mg/d, measured by 7d weighed intake, completed a duplicate diet collection over 24h. Concurrently, a 24h urine collection, validated using p-aminobenzoic acid, was obtained and one timed spot plasma sample taken. Duplicate diets, complete urine collections and plasma samples were analysed for total genistein and daidzein using liquid chromatography–MS to determine PO intake. The potential for 24h urinary excretion and plasma PO concentrations to reflect dietary intake was investigated. Mean estimated and measured dietary PO intakes were 12·3 and 11·0mg/d respectively. The correlation between estimated intake and measured intake of PO was highly significant (r 0·98, P<0·001). Urinary excretion (24h) and plasma concentrations of PO were significantly related to measured dietary PO intake (r 0·97, P<0·001 and r 0·92, P<0·001 respectively). The relationship between 24h urinary PO excretion and timed plasma concentrations was also significant (r 0·99, P<0·001). These findings validate the PO database and indicate that 24h urinary excretion and timed plasma concentrations can be used as biomarkers of PO intake.