Talking about emotion and putting feelings into words has been hypothesized to regulate emotion in psychotherapy as well as in everyday conversation. However, the exact dynamics of how different strategies of verbalization regulate emotion and how these strategies are reflected in characteristics of the voice has received little scientific attention. In the present study, we showed emotional pictures to 30 participants and asked them to verbally admit or deny an emotional experience or a neutral fact concerning the picture in a simulated conversation. We used a 2 × 2 factorial design manipulating the focus (on emotion or facts) as well as the congruency (admitting or denying) of the verbal expression. Analyses of skin conductance response (SCR) and voice during the verbalization conditions revealed a main effect of the factor focus. SCR and pitch of the voice were lower during emotion compared to fact verbalization, indicating lower autonomic arousal. In contradiction to these physiological parameters, participants reported that fact verbalization was more effective in down-regulating their emotion than emotion verbalization. These subjective ratings, however, were in line with voice parameters associated with emotional valence. That is, voice intensity showed that fact verbalization reduced negative valence more than emotion verbalization. In sum, the results of our study provide evidence that emotion verbalization as compared to fact verbalization is an effective emotion regulation strategy. Moreover, based on the results of our study we propose that different verbalization strategies influence valence and arousal aspects of emotion selectively.