Working alliance has been shown to be important in influencing the outcome of therapy. Research evidence suggests that characteristics of both clients and therapists impact on the development of the working alliance. Although attachment theory is well researched, there is relatively limited research on the relationship between both therapist and client attachment style and the working alliance; traditionally, research has placed greater emphasis on client characteristics. The current study examines the extent to which both client and therapist self‐reported attachment styles are related to the working alliance. Thirty clients and 42 therapists were recruited from primary care psychology services. Thirty client–therapist dyads were examined. Participants completed self‐report measures of anxiety and depression, attachment style and working alliance at a single time point. Client and therapist attachment security were not independently related to working alliance. However, there was a significant association between therapist insecure attachment and alliance in more symptomatic clients. There was also some evidence that therapists and clients with oppositional attachment styles reported more favourable alliances. The study suggests that the relationship between therapist attachment style and alliance is not straightforward. It is likely that the complexity of clients' presenting problems, coupled with interaction between client–therapist attachment styles, influences the therapeutic alliance. Key Practitioner Message Therapist insecure attachment may negatively affect the therapeutic alliance in more symptomatic clients. It is important to consider the interaction between client and therapist attachment and how these interactions influence the therapeutic alliance. Therapists should be aware of their own personal attachment style and reflect on how this might manifest during the therapeutic process.