Background <p>The association between mean arterial pressure (MAP) and diabetes has been explored in some studies; however, the results have been controversial.</p> Methods <p>This longitudinal cohort study included 8,430 males and 7,034 females. The follow-up period lasted for 13 years, from 1994 to 2016. We categorized the participants according to the MAP quartiles. We investigated the association between MAP and type 2 diabetes mellitus (T2DM) and expressed it as hazard ratios (HRs) with 95% confidence intervals (CIs) after adjusting for potential confounders.</p> Results <p>After multivariate adjustment, we found that MAP was not a risk factor for incident T2DM using Cox regression analysis. The HR at the highest quartile of MAP was 1.10 (0.75, 1.61), <i>p</i> = 0.6363, compared to the lowest quartile of MAP. In the subgroup analysis, there was no interaction between MAP and the development of T2DM.</p> Conclusions <p>MAP was not associated with the incidence of T2DM in Japan.</p> Clinical trial number <p>Not applicable.</p>