Although scholars have highlighted the usefulness of attachment theory for psychotherapy (e.g., Bowlby, 1988; Holmes & Slade, 2018; Mallinckrodt, 2010), minimal empirical research exists examining the relationship between client attachment and therapist skills. In this study, we first investigated the factor structure of the therapist- and client-rated Helping Skills Measure (HSM; Hill & Kellems, 2002) for 5,830 psychodynamic psychotherapy sessions of 202 adult community clients working with 25 doctoral student therapists in a university clinic. The multilevel-confirmatory factor analysis supported a 3-factor structure (Exploration, Insight, Action), stable across time, at the session level in psychodynamic psychotherapy. Next, using a dynamic structural equation model for 592 sessions of 37 clients working with 6 therapists using both the HSM and the Experiences in Close Relationships-Short Form (Wei et al., 2007), we found a slight increase in exploration and insight skills as rated by therapists, but no significant change in client attachment dimensions over time. For the model using the therapist-rated HSM, we found significant and positive auto correlations for Anxiety, Avoidance, and Action, and a significant and positive cross-lagged correlation for Avoidance in one session predicting Action in the next session. For the model using the client-rated HSM, we found significant and positive auto correlations for Anxiety, Avoidance, and Exploration, and significant and negative cross-lagged correlations for Anxiety in one session predicting Exploration and Action in the next session. We did not find any significant cross-lagged correlations for therapist skills in one session predicting client attachment dimensions in the next session. We provide suggestions for practice and research, including training in attachment-informed therapy to improve therapist responsiveness and linking associations between client attachment and therapist skills to client outcome.