Objective. The object of our study was to examine the effects of physical activity in 59 elderly self-sufficient patients with MCI, comparing those with diabetes and those without diabetes, and estimating cognitive improvement by means of the MMSE test. Moreover, we measured the possible effect of aerobic activity on HbA1C. Materials and methods. We recruited 59 old patients, with mean age of 69.7 years, afferent to MOV.I.S Onlus association in Giarre, operating within the University of Catania. Participants were divided into two groups: Consisting of 31 elderly with diabetes; Consisting of 28 elderly without diabetes. Patients were tested with MMSE at the beginning of the exercise program and after six months. Moreover, in elderly patients with diabetes we estimated HbA1C at the beginning, after three months and after six months. Results. We used the Paired T test to compare results in patients with diabetes and patients without diabetes. The MMSE test score was 25,65 +/- 3,16 and 26,7 +/- 1,55 in patients with diabetes and in patients without diabetes, respectively. After six months of aerobic exercise the score was 26,54 +/- 3,14 (P = 0,6) in elderly with diabetes, while in elderly patients without diabetes it was 27,66 +/- 1,2 (P = 0,009). HbA1C, evaluated at the beginning, after three and six months, while showing a positive trend, did not reach statistical significance (P(T0 vs T3) = 0,78; P(T0 vs T6) = 0,138; P(T3 vs T6) = 0,076). Conclusions. Physical activity in elderly patients seems to reduce cognitive impairment. However, in patients with diabetes this result is not statistically significant, probably because of a poor vascular condition. Hypoglycemia, insulin resistance and vascular complications may be correlated with cognitive impairment. In the literature physical activity also acts on HbA1C, determining its reduction. The results we obtained, in apparent contradiction, may be due to a short period of observation. Therefore a longer study period will be needed for re-evaluation by the authors.