Observational studies, primarily of a case-control design, have shown an inverse association of fruit and vegetable consumption with the risk of stomach cancer, a finding tentatively attributed to anti-oxidant vitamins. Ensuing randomized-intervention trials of these vitamins, however, have been mostly negative. Therefore, the seemingly protective effect of fruit and vegetables in case-control studies is suspected to be influenced by the information bias inherent in the retrospective assessment of exposure, particularly since pre-conceptions about the wholesome effects of these foods are common among the public. Our aim was to examine the association of fruit and vegetable intake with the risk of stomach cancer in a prospective cohort study. Fruit and vegetable consumption was assessed in 1967 in 11,546 individuals in the Swedish Twin Registry, along with a wide range of potentially confounding factors. Complete follow-up through 1992 was attained through record linkage to the National Cancer and Death Registers. The relative risk of stomach cancer was estimated in proportional hazards models, with confidence intervals (CIs) adjusted for correlated outcomes. The risk of stomach cancer was inversely related to fruit and vegetable consumption. Controlling for potentially confounding factors, the relative risk among subjects with the lowest compared to those with the highest intake was 5.5 (95% CI 1.7-18.3) with a statistically significant dose-risk trend (p < 0.05). Our results indicate that information bias is not likely to explain the discrepancy between the results of observational studies and of randomized-intervention trials.