Objectives: Inpatient psychotherapy includes the patient's manifold contacts with different therapists, nurses and fellow patients. The present study investigated the association between these multiple relationships and therapy outcome. Methods: Pre-post-measures of symptom load (Brief Symptom Inventory), interpersonal problems (IIP) and self-efficacy (SEB) were used to define three groups with positive (N = 129), unchanged (N = 44) or negative (N = 40) outcome. These groups were compared 1) by their alliance with an individual therapist, their relationship to the therapeutic team, their experience of cohesion and climate concerning fellow patients in the ward (measured weekly by the Stationserfahrungsbogen" SEB), and 2) by their differences in mean correlations between the courses of relationship experiences and symptom load. Results and conclusions: Cohesion and relationship to the therapeutic team were not associated with therapy outcome. Therapeutic alliance with the individual therapist and climate among fellow patients turned out to be moderate indicators of the therapeutic outcome. It is recommended to include these process parameters systematically into the process diagnostics of inpatient psychotherapy.