Objective: The purpose of this study was to examine the psychometric properties of the Body Image-Acceptance and Action Questionnaire in a severe eating disorder population, as previous validation has occurred only with nonclinical samples. Method: Data on body image psychological flexibility, general psychological flexibility, eating disorder severity, and other related constructs were gathered from 72 adolescent and 60 adult female, residential patients diagnosed with an eating disorder. Psychometrics were examined through the use of exploratory and confirmatory factor analyses, Cronbach's alpha, correlations, and hierarchical multiple regressions to assess model fit, reliability, and validity. Results: The BI-AAQ demonstrated excellent convergent, discriminant, and incremental validity as well as excellent internal reliability, however, factor analyses resulted in overall poor model fit. Removal of item 6 from the BI-AAQ resulted in improved psychometric properties in all regards, yet still demonstrated overall poor model fit. Discussion: This study suggests that the BI-AAQ is psychometrically sound in many areas and provides some clinical utility; however, it may be somewhat problematic when used in severe eating disorder populations. When using the measure in clinical settings, removal of item 6 is recommend. Recommendations for future measurement and utilization of body image flexibility are discussed.