Though there are extensive data on diabetic retinopathy, neuropathy and peripheral vascular disease, there are not enough data on diabetic stroke. Present study was carried out to evaluate distinctive feature of diabetic and non-diabetic stroke and to compare early mortality between two groups. Four hundred and fifty consecutive stroke cases categorised into group I of 171 known diabetic and 279 non-diabetics. Reactive hyperglycaemia seen in cases with HbA1c below 7.0% were put into group II. Patients were evaluated clinically. Evidence of peripheral vascular disease, retinopathy and other target organ damage were assessed. Glycaemic parameters monitored include fasting plasma glucose (FPG), random plasma glucose (RPG) and HbA1c. Total serum cholesterol was taken as the main risk factor in both groups though lipid profile was done in all cases. CT scan was done in all cases and a repeat one when needed. Out of 450 cases 171 were diabetic and 279 cases were non-diabetic. There was female preponderance in diabetic. Average age of diabetic stroke was lower than non-diabetic, 51.2 years as against 67 years. History of previous stroke was higher in diabetic (15.8%) than non-diabetic (5.7%). There was higher incidence of transient ischaemic attack in diabetic (22.8%) than non-diabetic (7.5%), which was highly significant (p<0.001). Hypertension and hypercholesterolaemia (> 175 mg/dl) were two important risk factors in 70.9% and 30.9% respectively in group I compared to 47.6% and 21.1 % respectively in group II, both of which are significant (p<0.001). Ischaemic stroke were higher in group I(69%) as compared to group II (45.8%) which was significant (p<0.001). Lacunar infarct were more in group I (73.7%) than group II (61.7%). Haemorrhagic stroke was higher in group II (52.7%) than in group I (30.4%). Apart from recent event CT scan showed evidence of old lacunar infarct in 36.8% cases of group I compared to 21.1% of group II. During follow-up mortality within 4 weeks was higher in haemorrhagic stroke of group I (55.8%) compared to 49.6% in group II which is significant (p<0.05). For ischaemic stroke mortality in group I was 26.3% compared to 14.8% in group II which is very significant (p<0.001). Out of total mortality in group I, 35.08% had high HbA1c. Increased risk of stroke in diabetic is probably related to hypertension and lipid abnormalities. Increased mortality from abnormal glycaemic control needs to be evaluated further.