Confirmatory factor analysis (CFA) and exploratory and factor analysis (EFA) aimed to determine the optimum Autism Spectrum Quotient--Children (AQ-Child) model. Initial CFA of parent ratings of the AQ-Child for 404 clinic-referred children with ADHD, aged between 4 and 11 years, revealed mixed/moderate support for the implied AQ-Child five-factor model and the past statistically supported four-factor model (Auyeung et al., J Autism Dev Disord 38:1230-1240, 2008). Interestingly, EFA findings indicated most support for a four-factor model, with factors reflecting "mind-reading", "social skills", "attention to details", and "imagination". The items loading in these factors were different from those proposed originally for similar factors (Auyeung et al., J Autism Dev Disord 38:1230-1240, 2008). The factors in the model showed acceptable internal consistency-reliability and discriminant validity. Clinical and research implications are discussed.