Low heart rate variability (HRV) is associated with an increased susceptibility to ventricular fibrillation (VF). Exercise training can increase HRV (an index of cardiac vagal regulation) and could, thereby, decrease the risk for VF. To test this hypothesis, a 2-min coronary occlusion was made during the last min of a 18-min submaximal exercise test in dogs with healed myocardial infarctions; 20 had VF (susceptible), and 13 did not (resistant). The dogs then received either a 10-wk exercise program (susceptible, n = 9; resistant, n = 8) or an equivalent sedentary period (susceptible, n = 11; resistant, n = 5). HRV was evaluated at rest, during exercise, and during a 2-min occlusion at rest and before and after the 10-wk period. Pretraining, the occlusion provoked significantly (P < 0.01) greater increases in HR (susceptible, 54.9 ± 8.3 vs. resistant, 25.0 ± 6.1 beats/min) and greater reductions in HRV (susceptible, −6.3 ± 0.3 vs. resistant, −2.8 ± 0.8 ln ms²) in the susceptible dogs compared with the resistant animals. Similar response differences between susceptible and resistant dogs were noted during submaximal exercise. Training significantly reduced the HR and HRV responses to the occlusion (HR, 17.9 ± 11.5 beats/min; HRV, −1.2 ± 0.8, ln ms²) in the susceptible dogs; similar response reductions were noted during exercise. In contrast, these variables were not altered in the sedentary susceptible dogs. Posttraining, VF could no longer be induced in the susceptible dogs, whereas four sedentary susceptible dogs died during the 10-wk control period, and the remaining seven animals still had VF when tested. Atropine decreased HRV but only induced VF in one of eight trained susceptible dogs. Thus exercise training increased cardiac vagal activity, which was not solely responsible for the training-induced VF protection.