Although the large majority of patients with end-stage renal disease are hypertensive, blood pressure levels are insufficiently controlled despite the complex pharmacological treatment they receive. Hypertension is known to be associated with cardiovascular complications, but its impact on mortality is still controversial. The paradox of “reverse epidemiology”, i.e. better survival of dialysis patients with high blood pressure levels, is common to other patients, such as those with heart failure, and is linked to the left ventricular abnormalities that are found in many hemodialysis (HD) patients. Moreover, unlike epidemiological studies on the general population, observational surveys on HD patients have been performed for shorter periods of time. After correction for confounders and over the long-term, hypertension seems to be an independent risk factor for mortality in HD patients and is multi-factorial in origin. Several causes are recognized: volume overload, sympathetic hyperactivity, activation of the renin angiotensin system, vascular calcification and erythropoietin treatment. The optimal blood pressure target seems to be pre-dialysis values of 140/90 mmHg, or the lowest possible values which are well tolerated during the dialysis session. The best method for assessing blood pressure levels remains to be established, because of oscillations in relation to the dialysis session or the inter-dialysis period. Home blood pressure measurement seems to be as sensitive as ambulatory blood pressure monitoring; alternatively the average of 12 predialysis measurements has been used. Drug treatment is unsatisfactory, and difficult kinetics and side effects are often reported. Several non pharmacological options are available, such as modifications to the hemodialysis schedule, increased HD efficiency, and use of biocompatible membranes, besides dietary salt and fluid restriction. A correct approach to HD can efficiently reduce the incidence of hypertension, cardiovascular complications and mortality rate in these patients.