Patients with congestive heart failure reportedly show a poor correlation between cardiac function and exercise tolerance. Recent studies have demonstrated that skeletal muscle is the main factor that limits exercise tolerance. However, the relationship between high-energy phosphate metabolism in skeletal muscle and exercise tolerance has not been well defined. Exercise capacity was assessed in 35 subjects with myocardial infarction in terms of peak oxygen consumption (peak VO2) during treadmill exercise with an analysis of expired gases. On the same day, changes in high-energy phosphates in finger flexor muscle during handgrip exercise were measured by magnetic resonance spectrometry. Phosphocreatine (PCr) utilization and the decrease in pH during handgrip exercise were significantly greater in patients with a poor exercise capacity and their time constant of recovery of PCr was prolonged. The ratios of PCr/Pi (Pi: inorganic phosphate) and PCr/HMPA (HMPA: hexamethylphosphoramide) during exercise and the time constant of recovery of PCr were significantly correlated with peak VO2. These results suggest that skeletal muscle metabolism governs exercise tolerance.