Ventricular arrhythmias, a leading cause of sudden cardiac death, can be triggered by cardiomyocyte early afterdepolarizations (EADs). EADs can result from an abnormal late activation of L-type Ca2+ channels (LTCCs). Current LTCC blockers (class IV antiarrhythmics), while effective at suppressing EADs, block both early and late components of I-Ca,I-L, compromising inotropy. However, computational studies have recently demonstrated that selective reduction of late I-Ca,I-L (Ca2+ influx during late phases of the action potential) is sufficient to potently suppress EADs, suggesting that effective antiarrhythmic action can be achieved without blocking the early peak I-Ca,I-L, which is essential for proper excitation-contraction coupling. We tested this new strategy using a purine analogue, roscovitine, which reduces late I-Ca,I-L with minimal effect on peak current. Scaling our investigation from a human Ca(V)1.2 channel clone to rabbit ventricular myocytes and rat and rabbit perfused hearts, we demonstrate that (1) roscovitine selectively reduces I-Ca,I-L noninactivating component in a human Ca(V)1.2 channel clone and in ventricular myocytes native current, (2) the pharmacological reduction of late I-Ca,I-L suppresses EADs and EATs (early after Ca2+ transients) induced by oxidative stress and hypokalemia in isolated myocytes, largely preserving cell shortening and normal Ca2+ transient, and (3) late I-Ca,I-L reduction prevents/suppresses ventricular tachycardia/fibrillation in ex vivo rabbit and rat hearts subjected to hypokalemia and/or oxidative stress. These results support the value of an antiarrhythmic strategy based on the selective reduction of late I-Ca,I-L to suppress EAD-mediated arrhythmias. Antiarrhythmic therapies based on this idea would modify the gating properties of Ca(V)1.2 channels rather than blocking their pore, largely preserving contractility.