To investigate the determinants of the therapeutic working relationship and better understand its intrapersonal and interpersonal nature, this study investigated therapist characteristics as predictors of the formation and development of patient‐rated and therapist‐rated working alliances within a clinical trial of short‐term versus long‐term therapies. Short‐term (solution‐focused and short‐term psychodynamic) and long‐term (long‐term psychodynamic therapy and psychoanalysis) therapies were provided by 70 volunteering, experienced therapists to 333 patients suffering from depressive and/or anxiety disorders. Therapists' professional and personal characteristics, measured prior to the start of the treatments, were assessed with the comprehensive self‐report instrument, Development of Psychotherapists Common Core Questionnaire. The Working Alliance Inventory was rated by both therapists and patients at the third session and at the 7 months' follow‐up point from the initiation of therapy. Therapists' self‐rated basic interpersonal skills were found to predict the formation of better patient‐rated alliances in both short‐term and long‐term therapies. Engaging, encouraging relational style fostered improvement of patients' working alliances especially in the course of short‐term therapies. However, it led to patient alliance deterioration in long‐term therapies, where constructive coping techniques proved more beneficial. Therapists' professional self‐confidence and work enjoyment, along with their self‐experiences in personal life, consistently predicted their alliances, but were less salient for patient ratings of alliance. The divergence of therapist and patient viewpoints has implications for therapist training and supervision, as characteristics found detrimental or helpful for the working relationship rated from the perspective of one party may not be predictive of the other therapy participant's experience. Copyright © 2013 John Wiley & Sons, Ltd. Key Practitioner Message In both short‐term and long‐term therapies, patients form more positively evaluated working alliances with those therapists who experience their overall basic relational skills as better. However, as patients' experience of alliance also develops during the course of therapy, it seems that briefer treatments may benefit more from a therapist's engaging and affirming relational style than long‐term therapy.Practitioners' professional lack of confidence and enjoyment in therapy work predict poorer working alliances as rated by themselves, but therapists generally seem to be able to contain these difficult experiences, so that they do not predict worse patient‐rated alliances.Therapists' self‐experiences in their close personal relationships, such as how open, forceful or private they experience themselves, predict how they will experience the working relationships with their patients.Practitioners should be aware that how they evaluate their working relationships with specific patients may be notably influenced by their overall experiences of their professional and private selves and also that many of these self‐experienced qualities have less strong or direct bearing on their patients' experiences.