Monitoring of fruit and vegetable (F&V) intake is fraught with difficulties. Available dietary assessment methods are associated with considerable error and use of biomarkers offers an attractive alternative⁽Reference Pollard, Wild and White¹⁾. Few studies to date have examined the use of plasma biomarkers to monitor or predict F&V intake of volunteers consuming a wide range of intake from both habitual F&V and manipulated diets. This study combined plasma vitamin C and carotenoid concentrations with Ferric Reducing Antioxidant Power (FRAP) as an integrated biomarker of F&V intake and compared the predictive powers of each single and integrated biomarker for F&V intake. Data from a randomized, controlled, dietary intervention study (FLAVURS) (n = 154) in which the test groups had observed sequential increases of 2·3, 3·2 and 4·2 portions of F&V every 6-wk across an 18-wk period was used in this study. A modified integrated plasma biomarker was devised which included plasma vitamin C, total carotenoids and FRAP values, and this gave a better correlation with F&V intake (r = 0·516, P < 0·001) than the individual biomarkers (r = 0·332, P < 0·001; r = 0·417, P < 0·001; r = 0·136, P = 0·099 respectively). Inclusion of urinary potassium concentration did not significantly improve the correlation. The modified integrated plasma biomarker more accurately predicted F&V intake to within 2 portions of the actual intake in 54·3 ± 4·9% of the population compared with plasma carotenoid concentration (48·3 ± 11·3%), although this difference did not reach statistical significance (P > 0·05). Either plasma carotenoid concentration or the integrated biomarker could be used to distinguish high and low F&V consumers.