The aim of this study was to classify antiarrhythmic drugs based on their effectiveness on 6 in vivo arrhythmia models, mainly using dogs. The models were produced by two-stage coronary ligation, digitalis, halothane-adrenaline, programmed electrical stimulation in old myocardial infarction dogs, coronary artery occlusion/reperfusion, or chronic atrioventricular block. Na(+)-channel-blocking drugs suppressed two-stage coronary ligation and digitalis arrhythmias. Ca(2+)-channel blockers and beta-blockers suppressed halothane-adrenaline arrhythmia. Positive inotropic drugs aggravated halothane-adrenaline arrhythmia, but did not aggravate digitalis arrhythmia. K(+)-channel blockers suppressed programmed electrical stimulation induced arrhythmia, but induced torsades de pointes type arrhythmia in chronic atrioventricular block dogs and aggravated halothane-adrenaline arrhythmia. Na(+)/H(+)-exchange blockers suppressed coronary artery occlusion/reperfusion arrhythmias. This classification may be useful for predicting the clinical effectiveness in the preclinical stage of drug development.