Objectives To examine the relationship between improvements in adolescent ratings of therapeutic alliance and reductions in posttraumatic stress disorder (PTSD) severity over time among adolescent girls during prolonged exposure therapy for adolescents (PE‐A) versus client‐centered therapy (CCT), as well as to examine differences in changes in alliance between treatment groups. Method A total of 61 adolescent girls (aged 13–18 years) with sexual assault‐related PTSD received PE‐A or CCT in a randomized controlled trial. Participants rated alliance at session 3, midtreatment, and posttreatment. Results The rate of improvement in adolescent‐rated alliance was greater in PE‐A than CCT over the course of treatment. In addition, improvement in adolescent‐rated alliance significantly contributed to improvements in PTSD (regardless of treatment condition), but not vice versa. Conclusions Contrary to beliefs that trauma‐focused treatments fail to establish strong therapeutic alliance in sexually abused adolescents, improvement in adolescent ratings of alliance were greater in PE‐A compared to CCT, and improvements in adolescent‐rated alliance were significantly associated with better treatment outcome across both types of treatments.