Left ventricular (LV) diastolic dysfunction is observed frequently in patients with type 2 diabetes; however, few studies have focused on the effect of the Rho-associated kinase inhibitor fasudil on cardiac performance in humans. We conducted a prospective pilot study to assess the impact of fasudil on LV diastolic function in patients with diabetes without systolic dysfunction. Two hundred and fifty eligible patients with type 2 diabetes (149 men [61.3 %] and 94 women [38.7 %]) with a mean age of 57.2 years were randomly assigned to fasudil (n = 122, 30 mg intravenously twice a day for 14 days) or placebo (n = 121) groups. Echocardiographic variables were measured at the baseline and 1 month after the intervention. Compared with the placebo group, the fasudil group showed a significant decrease in diastolic blood pressure and in the peak of late diastolic transmitral flow (Am) (P < 0.05 for both). Deceleration time (DT), isovolumic relaxation time (IVRT), the peak of early diastolic annular velocity (e′), the peak of late diastolic annular velocity, and E/e′ also exhibited a significant improvement (all, P < 0.05) after fasudil administration. Furthermore, the Em/Am ratio and IVRT, DT, and E/e′ values recorded after fasudil treatment in the subgroup with impaired LV relaxation significantly differed from the corresponding values in the subgroup with normal LV relaxation (all, P < 0.05). Fasudil improves short-term echocardiographic parameters of LV diastolic function in patients with type 2 diabetes with preserved left ventricular ejection fraction.