Objective: This study investigates the therapeutic process of one person receiving forty sessions of Affect Phobia Therapy (APT; McCullough, 1997). APT postulates systematic desensitization of affect phobia to be essential for affect restructuring, leading to therapeutic healing. The four process variables hypothesized to be essential in this process were examined and include; Insight, Motivation, Activating affects and Inhibiting affects. Change in process variables were analyzed between-sessions across the course of therapy, and potential sequential relationships between the variables were analyzed within-sessions. Method: The data analyzed was from a Randomized Controlled Trial of brief Cognitive Therapy and APT (Svartberg et al. 2004). The participant was diagnosed with Cluster C personality disorder and assigned to the APT condition. Outcome assessments consisted of SCL-90-R, IIP-127 and MCMI-III. The outcome data at treatment termination indicated some symptom reduction, but suboptimal therapeutic effects regarding personality-and interpersonal problems. Two-year follow-up showed significant change on all outcome measures. All videotaped sessions were observed and rated on a quantitative process measure, the Achievement of Therapeutic Objectives Scale (ATOS; McCullough et al., 2003). Between-session analyses were conducted using linear regression models. Within-session analyses were performed using ARIMA time series models. Additional qualitative descriptions were included to illustrate the basis of ATOS scorings and used for discussing the therapeutic process and the proposed theory of change. Results: The predicted increase in Insight, Motivation and Activating affect, and decrease in Inhibiting affects across therapy were not confirmed in this study. Neither were the predicted within-session temporal relationships between the process variables found. Only one temporal relationship between the process variables was found and indicated that Insight in one segment predicted Activating affect in the next. Conclusion: The suboptimal change rated on the process variables was indicative of a less than optimal therapy course. These findings may illuminate the patient’s outcome scores at treatment termination. The outcome at two-year follow-up could not be readily predicted by the ATOS process-scores, but may be understood in light of other theories of therapeutic change.