Abstract Several case-control studies have reported that patients with diabetes mellitus (DM) had a higher prevalence of Helicobacter pylori infection than those without DM, but these findings remain equivocal. Additionally, there are few studies examining associations between East Asian CagA-positive H. pylori and DM. This cross-sectional study aimed to investigate whether H. pylori infection was a possible risk factor for DM in a general Japanese population. The study included 5165 subjects (1467 men, 3698 women) aged 35–69 years from the Daiko Study, part of the Japan Multi-Institutional Collaborative Cohort Study. A urinary anti- H. pylori antibody was used to detect H. pylori infection. The medical history of physician-diagnosed DM was confirmed by self-administered questionnaire. The odds ratios (ORs) and their 95 % confidence intervals (CIs) for DM (current and former) were calculated using unconditional logistic regression analysis, adjusting for age, sex, educational status, alcohol use, smoking status, body mass index, energy intake, and physical activity. The prevalence of DM was 4.6 % (95 % CI 3.7–5.6 %) among 1878 participants with H. pylori infection and 3.2 % (2.6–3.8 %) among 3287 without the infection ( p = 0.009). The crude, age-adjusted, and multivariate-adjusted ORs for DM in those with the infection relative to those without were 1.47 (95 % CI 1.10–1.97), 1.02 (0.76–1.38), and 0.97 (0.71–1.32), respectively. We found a significantly higher DM prevalence among those with H. pylori infection than among those without. However, almost all the difference in prevalence could be explained by the older age of those infected. Our findings did not support an association between H. pylori infection and DM.