Objective: Investigate the possible predictors of stroke, according to a five-year follow-up in hypertensive patients. Design and method: Eighty essential hypertensive patients aged between 30–59 years were observed within 5 years. General clinical investigation, 24-h ABPM and brain MRI were performed in all patients at the onset of the study and after 5 years observation. Presence of increased subarachnoidal spaces (SS) and lateral ventricles (LV) of a brain, periventricular leucoaraiosis (PL), lacuna infarcts (LI) and focal white matter lesions (FWML) were evaluated in all patients. Results: 13,7% of patients during 5 years experienced stroke. This patients (I group - 13,7 %) were compared to group of hypertensive patients without stroke (II group – 86,3 %) with regard to baseline data. At patients with insults on initial brain tomograms authentically were more often registered FWML (75% vs 42%; chi square = 4,49, p = 0,046), LA (75% vs 44%; chi square = 5,04, p = 0,024) and LI (50% vs 6%;chi square = 17,04, p = 0,000). At baseline, 24-h systolic (170.7 ± 18.09 vs 149.8 ± 15.02, p = 0,000), diastolic (106.8 ± 9,70 vs 97.9 ± 10,73, p = 0.003) blood pressure and the number of night-pikkers (45.5% vs 15.9%; chi square = 5,14, p = 0,025) were higher at patients of I group. In multiple regression model including clinical data, baseline MRI signs and BP parameters, daytime systolic blood pressure (R = 0,38, p = 0,002), daytime systolic BP load (R = 0,41; p = 0.002), LA (R = 0,42; p = 0,025) and FWML (R = 0,39; p = 0,012) were best predictors of stroke. Significance of Multiple linear regression model: R = 0,53; R2 = 0,28; p = 0,0007. Conclusions: Our study demonstrates that systolic BP, focal white matter lesions and leucoaraiosis are independent predictors of stroke in hypertensive patients. Thus, ABPM and brain MRI early may significantly improve risk stratification in hypertensive patients.