Dysfunctional beliefs and cognitive biases play an important role in the formation and maintenance of depression according to current cognitive theories. These refer to 'thinking errors' as conceptualized in cognitive behavioural interventions (e. g., over-generalization), but also to depression-specific, cognitive biases as identified by basic research (e. g., mood-congruent recall). The latter is hardly picked up by existing treatment approaches. The Metacognitive Training for Depression (D-MCT) was designed to fill this gap. Aim of the current pilot study was to investigate the feasibility, acceptance, as well as outcome of the treatment program. One hundred and four patients with depression participated in an open-label pilot study. Per protocol analysis included 72 patients. Sociodemographic data, depressive symptoms, self-esteem, rumination, as well as general metacognitive and depression-specific cognitive biases were assessed by a range of self-rating instruments. Subjective appraisal of the training as well as potential overlap with other treatments regarding content was additionally inquired. Analyses revealed a significant decrease of depressive symptoms with a medium effect (Cohen's d = 0.73). Moreover, cognitive biases and rumination were significantly reduced and self-esteem was increased at the same time (with effect sizes between d = 0.32 and 0.64). According to the results of the present pilot study, the D-MCT represents a promising new group treatment in terms of feasibility, acceptance, and outcome.