The extensive reliance on symptoms for the study of psychotherapy is often criticized. In this study we examined whether the subjective sense of mental pain predicts psychotherapy process and outcome, above and beyond the effect of symptomatic distress. Outpatients (n = 53) treated in a psychiatric hospital completed measures of mental pain intensity and tolerance, symptomatic distress, and session climate at pretreatment and posttreatment. Multilevel modeling was utilized to assess the predictive effect of mental pain, while controlling baseline symptomatic distress. Patients with high mental pain at baseline showed significant reductions in distress, while patients with low mental pain showed no significant improvement. Moreover, low mental pain and high mental pain tolerance predicted decreases in session smoothness. Mental pain can serve as a predictive marker for psychotherapy process and outcome, and complement the reliance on symptomatic distress in psychotherapy research.