Objectives Recent studies have shown a bidirectional association between diabetes and periodontal disease. However, the longitudinal association between periodontal pocket depth and new onset of diabetes remains controversial. Therefore, this study aimed to examine the association between periodontal pocket depth and new onset of diabetes using dental check-up data in a community-based population.Methods The study comprised 5,163 participants aged≥20 years (mean age±standard deviation, 57.4±13.9 years; women, 66.3%) who underwent dental check-ups between April 2016 and March 2019 in a Tokyo ward. We classified the participants into three groups using community periodontal index codes from dental check-ups: healthy periodontal pocket group, periodontal pocket 4-5 mm group, and periodontal pocket ≥6 mm group. The participants were followed until the end of March 2020. Additionally, data on individuals who developed diabetes in the same ward were acquired from the National Health Insurance and Latter-Stage Older Persons Health Care System using ICD-10 codes. As the study outcome, we identified individuals who developed diabetes after the dental check-up date. Diabetes incidence rates were compared using the log-rank test and Cox regression analysis. Additionally, we performed a sensitivity analysis using a similar framework.Results The log-rank test showed that the cumulative incidence of diabetes between the three groups was significantly different (P<0.01). A Cox regression analysis model adjusted for sex, age, smoking habits, number of teeth present, and oral hygiene status showed that the diabetes development hazard ratio (HR) for the periodontal pocket ≥6 mm group was 1.44 (95% confidence interval [CI]; 1.04-2.00) when compared with that of the healthy periodontal pocket group. In the sensitivity analysis of individuals aged ≥40 years, the HR for the periodontal pocket ≥6 mm group was 1.55 (95% CI; 1.11-2.16) when compared with that of the healthy periodontal pocket group. Similarly, among men aged ≥40 years, the HR for the periodontal pocket ≥6 mm group was 1.72 (95% CI; 1.04-2.85) when compared with that of the healthy periodontal pocket group. However, no significant association between new onset of diabetes and periodontal pocket depth was found for women aged ≥40 years (HR=1.39, 95% CI; 0.89-2.18).Conclusion This The study suggests a longitudinal association between periodontal pocket depth and new onset of diabetes. The association was particularly pronounced in men aged ≥40 years. Therefore, men aged <40 years should maintain good oral health through appropriate dental health guidance to prevent the future development of diabetes.