The initiation mechanism of drug-induced arrhythmia has been explained by EAD occurring under prolonged action potential (AP) due to I_Kr block. Thus, I_Kr block potency and AP prolongation have been used in drug screening to predict drug-induced arrhythmia. However, even if I_Kr blockers cause the prolongation in the same level, they may have different risks for drug-induced arrhythmia. For example, amiodarone prolongs AP but does not induce drug-induced arrhythmias and is considered as a relatively safe drug in clinical practice. On the other hand, terfenadine and bepridil prolong APs and induce drug-induced arrhythmia. This different risk among I_Kr blockers implies that additional mechanisms may contribute to induction and suppression of EAD. In the present study, we studied how different block kinetics of non-selective I_Kr blockers on L-type Ca²⁺ current (I_CaL) affects occurrence of EAD. By using a mathematical model of the human ventricular action potential, we found that the different drug block kinetics on I_CaL can account for the different occurrence of EAD. Voltage-independent I_CaL block suppressed EAD under prolonged AP. The I_CaL block model of amiodarone also suppressed EAD effectively. In contrast, the I_CaL block models of terfenadine and bepridil increased the occurrence of EAD. These results suggest that, to predict drug-induced arrhythmia, not only I_Kr block and AP prolongation but also voltage-dependent property of I_CaL block should be checked.