•Children in community sample had a lower prevalence of attachment disorders compared to children in clinic and foster care.•Attachment disorder was associated with psychopathological symptoms and higher prevalence of mental disorders.•Attachment disorder was associated with lower cognitive and language abilities.•A high percentage (37.1%) of children fulfilled the diagnostic criteria for both types of attachment disorders.•Our findings corroborate previous literature suggesting that attachment representation is distinct from attachment disorder. Currently, attachment quality and attachment disorder exist in parallel, but the mutual association is still insufficiently clarified. For policy makers and clinical experts, it can be difficult to differentiate between these constructs, but the distinction is crucial to develop mental-health services and effective treatment concepts. We aimed to investigate the association between attachment representations (AR) and attachment disorders (AD), including Reactive Attachment Disorder (RAD) and Disinhibited Social Engagement Disorder (DSED) in children aged between 5 and 9. A total of 135 children aged between 5 and 9 years (M=7.17 years, SD=1.40, 63% male) and their primary caregivers participated in the study. Children were interviewed with the story stem method to assess AR, and the primary caregiver completed diagnostic interviews and questionnaires on mental disorders, AD, emotional and behavioral problems, and intelligence and development. The prevalence of AR in children with AD was 28.6% for the ‘secure’ form of AR, 17.1% for the ‘insecure-avoidant’ form, 25.7% for the ‘insecure-ambivalent’ form, and 28.6% for the ‘disorganized’ form. Prevalences of the various AR forms did not differ statistically significantly, indicating that AR is conceptionally distinct from AD. Children with disorganized attachment scored significantly lower on language and intelligence skills than children with secure attachment. AD was significantly associated with a higher number of comorbidities, emotional and behavioral problems, and lower language skills. Longitudinal studies using standardized assessment instruments are needed to systematically provide comparable and reliable empirical findings to improve current understanding of AR and AD as well as their etiological models.