Hypervolemia, present in at least 70% of patients with decompensated heart failure, results in renal dysfunction due to increased renal venous pressure, impaired renal autoregulation, and decreased renal blood flow that are associated with increased morbidity and mortality. Loop diuretics, widely used in congested patients, result in the production of hypotonic urine and neurohormonal activation. In contrast, ultrafiltration (UF) removes isotonic fluid without increasing renin secretion by the macula densa. Simplified devices that permit us to perform UF with peripheral venous access, adjustable blood flows, and small extracorporeal blood volumes make this therapy feasible at most hospitals and in less acute care settings. Conflicting results on the effects of UF in heart failure patients underscore the challenges of patient selection and choice of fluid removal rates. Unfavorable outcomes in patients undergoing UF in the midst of cardiorenal syndrome type 1 are in contrast with the sustained benefits of UF initiated before unsuccessful use of high-dose intravenous (IV) diuretics. UF rates should be based on a precise knowledge of the degree of hypervolemia and careful assessment of blood volume changes, so that extracellular fluid gradually refills the intravascular space and volume depletion is avoided. Poor outcomes are likely to occur if fluid removal rates are not tailored to individual patients' clinical characteristics. A large trial is ongoing to determine if a strategy of early UF, initiated before renal function is worsened by other therapies, is superior to IV diuretics in reducing 90-day heart-failure-related hospitalizations in patients with pulmonary and systemic congestion.