Experimental models of heart failure can be used to address specific questions not easily answered in patients, but no single model can reproduce exactly any of the clinical syndromes of heart failure since these are dominated by fatigue and breathlessness. Heart failure may be induced experimentally by pressure loading, volume loading, myocardial infarction, or by the creation of other disease states within the myocardium. Pressure loading may be especially useful in the study of ventricular hypertrophy, cellular derangements and vascular changes. Volume loading may be useful when examining the pathogenesis of hormone and electrolyte disturbances. Models of myocardial infarction or destruction are likely to be the most suitable for assessing novel therapy provided that peripheral reflexes are maintained. Experimental cardiomyopathy can provide an important means of identifying pathological subcellular mechanisms. They may be of use in the evaluation of vasodilator drugs but caution should be exercised in the study of inotropic agents. Any one model may be useful if it permits study of a single factor or variable in isolation or at a time when information is not obtainable from patients. For greatest clinical relevance, studies should be made in conscious animals with intact reflexes.