Many guidelines recommend increases in potassium intake. However, the relationship of dietary potassium intake with incident cardiovascular disease (CVD) has not been examined in those with type 2 diabetes (T2DM), including sodium acting antagonistically with potassium. We investigated these relationships in Japanese patients with T2DM. The investigation was part of the JDCP study, a nationwide prospective study begun in 2007. Analysed were 1477 persons with T2DM, 40-75 years of age, who completed a brief-type, self-administered Diet History Questionnaire at baseline. Primary outcome was a CVD event during the follow-up median 7 years (3.9-8.1 years). Hazard ratios (HRs) for CVD were estimated by Cox regression adjusted for confounders of daily potassium intake categorized by tertiles. Tertiles of sodium intake were also analysed. Mean daily potassium intake in tertiles was 1877, 2627 and 3532 mg, respectively, and significant associations were not shown between potassium intake and incidence of CVD. When HRs for CVD were stratified for potassium intake in tertiles (reference group, bottom tertile) and sodium intake (reference group, bottom tertile), potassium intake in the bottom tertile and sodium intake in the second and top tertiles were associated with significantly elevated HR for CVD (2.79 [1.02-7.63] and 3.92 [1.30-11.79], respectively). Low potassium intake in conjunction with high sodium intake was significantly associated with increased incident CVD in persons with T2DM. However, CVD incidence was not related to high potassium intake, regardless of sodium intake.