Many epidemiological studies have examined the associations of nutrients and foods with risk of stomach cancer. Few studies have addressed the dietary patterns that reflect various components of food consumption and their interactions. We identified major four dietary patterns; "vegetable and fruit", "Western breakfast", "meat", and "rice/snack" with principal component analysis using food consumption questionnaire in a prospective study of 5,765 middle-aged male workers in Tokyo. After 10 years of follow-up between September 1988 and August 1998, 84 incident cases of stomach cancer were documented. Using proportional hazards regression to estimate risk ratios, we found no clear association between each of dietary pattern and stomach cancer risk. After adjustment for age and other potential confounding factors and after exclusion of the cases diagnosed in first follow-up year, the risk ratio (RR) associated with high tertile compared to low tertile was 0.78(95%CI 0.42-1.44) for "vegetable and fruit" pattern and 0.71(95%CI 0.40-1.24) for "Western breakfast" pattern. The V-shaped associations between dietary patterns and stomach cancer risk were appeared in the "meat" (RR=1.00, 0.55, and 1.10) and the "rice/snack" (RR=1.00, 0.52, and 1.19) patterns, while the linear trend of these associations was statistically on borderline. The roll of overall dietary patterns in predicting stomach cancer risk requires further investigation.