Difficulties with emotion regulation are common in children with autism. Although interventions targeting emotion regulation show promise, children's individual treatment responses vary, and it is important to understand the factors that contribute to treatment change. The present study aimed to identify pre-treatment child characteristics and parent psychopathology that predict treatment response in a 10-week manualized cognitive behaviour therapy intervention for children with autism, 8-12 years of age. Exploratory best-subset regression analyses were first carried out to identify the optimal set of predictors. Logistic regressions were then conducted to determine whether these variables predicted reliable improvement. Outcome variables consisted of the lability/negativity and the emotion regulation subscales of the Emotion Regulation Checklist. Predictors included pre-treatment developmental, clinical, and parent psychopathology variables. Analyses revealed that youth who started the treatment with higher verbal reasoning, higher impairment in social motivation, and more anxious parents were more likely to show reliable improvements in emotion regulation. Youth who started the treatment with higher internalizing scores had lower odds of showing reliable improvement. Implications of our findings include facilitation of active involvement, avoidance of complex language, and the provision of additional supports. Further suggestions to inform clinical practice are discussed.