This study evaluated the diagnostic significance of a magnetic resonance imaging (MRI) based scoring model for identification of arrhythmogenic right ventricular cardiomyopathy (ARVC) in patients with MRI evidence of RV abnormalities. Fifty-three patients with RV myocardial abnormalities on MRI were divided into a group with ARVC 1 (n=17) and a group with other RV arrhythmias (n=37). Decision tree learning (DTL) and linear classification (based on a modified ARVC scoring model of major and minor criteria) were used to identify and assess MRI criterion information value, and to induce ARVC diagnostic rules. All major ARVC criteria were more frequent in the ARVC group. Among minor criteria regional RV hypokinesia, mild segmental RV dilatation, and prominent trabeculae were more frequent in the ARVC group while mild global RV dilatation was more frequent in the non-ARVC group. RV aneurysm achieved highest importance in ARVC diagnosis (predictive accuracy 76.8%). Better diagnostic accuracy (sensitivity 93.3%, specificity 89.5%) was achieved when the MRI score for the major and minor criteria reached threshold value of four: two major criteria, or one major and two minor, or four minor criteria. Combinations between major and minor criteria contributed to a statistically valid model for ARVC diagnosis.