Mental health conditions are more prevalent in autistic people than non-autistic people. The Strengths and Difficulties Questionnaire is a commonly used screening tool for emotional and behavioural difficulties in autistic children and adolescents. However, few studies have assessed the psychometric properties of the Strengths and Difficulties Questionnaire in populations of autistic people, especially measurement invariance, to examine whether the underlying latent structure is consistent across time or groups. Measurement invariance of the parent-reported Strengths and Difficulties Questionnaire in a nationally representative cohort study was examined: longitudinal invariance (autistic 11-, 14- and 17-year-olds) and group invariance (autistic and non-autistic 17-year-olds). Differential item functioning was explored in the autistic/non-autistic group analysis. The five-factor structure showed inadequate fit, especially in the autistic group; by extension, measurement invariance could not be established. Differential item functioning was observed for several conduct problems, prosocial behaviour and peer problems subscale items, as well as reverse-coded items. Poor structural validity and item-level invariance may confound findings from research within and between populations of autistic and non-autistic adolescents, as well as screening of mental health conditions in autistic adolescents. This research calls for the development and improvement of screening tools for emotional and behavioural difficulties in populations of autistic people while accounting for their heterogeneity.