The assessment of functional capacity in heart disease remains problematic, and it is unclear whether maximal exercise testing is physiologically reflective of the activities of daily living or the quality of life. We therefore employed a symptom-limited, self-paced walking protocol to assess the physical conditions of 41 Nigerian cardiac patients, with and without heart failure. The walking time, walking speed and distance as well as the energy expenditure (Kcal . min-1) were markedly reduced (P < 0.001) in patients with heart failure (n = 26) compared to the cardiac patients not in failure (n = 15). The double product corrected for exercise time (an index of myocardial oxygen use) was, however, significantly higher (P < 0.001) in the group with heart failure. Using multiple regression analysis, the parameters of self-paced walking capacity (distance, walking time, and speed) could reliably be predicted (r2 > 0.9) from age, body surface area, energy expenditure, and echocardiographic left ventricular dimension in the patients without heart failure. The presence of heart failure appeared to weaken the predictability of the regressions. A significant correlation was obtained between the self paced exercise time and the Bruce protocol treadmill time (r2 = 0.91, P = 0.004) in a subgroup of the patients with heart failure. Thus, the self-paced walking test is sensitive to changes in congestive heart failure and the exercise capacity can be predicted from age and biophysical parameters. The wider clinical usage of this modality, especially in frail patients, is hereby recommended.