The diminished fear reactivity is one of the most valid physiological findings in psychopathy research. In a fear conditioning paradigm, with faces as conditioned stimulus (CS) and electric shock as unconditioned stimulus (US), we investigated a sample of 14 high psychopathic violent offenders. Event related potentials, skin conductance responses (SCR) as well as subjective ratings of the CSs were collected. This study assessed to which extent the different facets of the psychopathy construct contribute to the fear conditioning deficits observed in psychopaths. Participants with high scores on the affective facet subscale of the Psychopathy Checklist-Revised (PCL-R) showed weaker conditioned fear responses and lower N100 amplitudes compared to low scorers. In contrast, high scorers on the affective facet rated the CS+ (paired) more negatively than low scorers regarding the CS- (unpaired). Regarding the P300, high scores on the interpersonal facet were associated with increased amplitudes to the CS+ compared to the CS-, while the opposed pattern was found with the antisocial facet. Both, the initial and terminal contingent negative variation indicated a divergent pattern: participants with pronounced interpersonal deficits, showed increased cortical negativity to the CS+ compared to the CS-, whereas a reversed CS+/CS- differentiation was found in offenders scoring high on the antisocial facet. The present study revealed that deficient fear conditioning in psychopathy was most pronounced in offenders with high scores on the affective facet. Event related potentials suggest that participants with distinct interpersonal deficits showed increased information processing, whereas the antisocial facet was linked to decreased attention and interest to the CS+. These data indicate that an approach to the facets of psychopathy can help to resolve ambiguous findings in psychopathy research and enables a more precise and useful description of this disorder.