Working alliance is considered an important determinant of outcome of psychotherapy. Patients with posttraumatic stress disorder (PTSD) following childhood abuse (CA-PTSD) may have challenges in building interpersonal relationships, including working alliance. Phase-based treatment provides an opportunity to strengthen alliance prior to trauma-focused treatment. This study aimed to compare the development of working alliance among patients with CA-PTSD in three variants of prolonged exposure (PE) therapy: standard PE, intensive PE (iPE), and skill training in affective and interpersonal regulation + prolonged exposure (STAIR + PE). We also examined the effect of alliance on treatment outcome and dropout. Self-reported PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders, fifth edition (Blevins et al., 2015) and patient-rated Working Alliance Inventory (Tracey & Kokotovic, 1989) were assessed in a clinical trial. We analyzed data from 138 adult patients (76.1% female; 42% non-Western). Analyses were performed using mixed-effects models. Patients established a satisfactory alliance early in treatment, which increased over time. For PE and STAIR + PE, a larger decrease in PTSD symptom severity was related to a higher alliance in the subsequent session, but not the other way around. In STAIR + PE, a higher alliance in Phase 1 was related to lower PTSD symptoms in Phase 2. In all conditions, a higher initial working alliance was related to a lower chance of treatment dropout. In the treatment of CA-PTSD, all three variants of prolonged exposure foster positive development of the working alliance. Across conditions, working alliance did not precede symptom decline. Therapists should strive for a strong alliance at the beginning of treatment as this reduces the likelihood of dropout. (PsycInfo Database Record (c) 2024 APA, all rights reserved).