Background Several studies suggest significant relationships between migration and autism spectrum disorder (ASD) but there are discrepant results. Given that no studies to date have included a pathological control group, the specificity of the results in ASD can be questioned. Aims To compare the migration experience (premigration, migratory trip, postmigration) in ASD and non-ASD pathological control groups, and study the relationships between migration and autism severity. Method Parents’ and grandparents’ migrant status was compared in 30 prepubertal boys with ASD and 30 prepubertal boys without ASD but with language disorders, using a questionnaire including Human Development Index (HDI)/Inequality-adjusted Human Development Index (IHDI) of native countries. Autism severity was assessed using the Child Autism Rating Scale, Autism Diagnostic Observation Schedule and Autism Diagnostic Interview-Revised scales. Results The parents’ and grandparents’ migrant status frequency did not differ between ASD and control groups and was not associated with autism severity. The HDI/IHDI values of native countries were significantly lower for parents and grandparents of children with ASD compared with the controls, especially for paternal grandparents. Furthermore, HDI/IDHI levels from the paternal line (father and especially paternal grandparents) were significantly negatively correlated with autism severity, particularly for social interaction impairments. Conclusions In this study, parents’ and/or grandparents’ migrant status did not discriminate ASD and pathological control groups and did not contribute either to autism severity. However, the HDI/IHDI results suggest that social adversity-related stress experienced in native countries, especially by paternal grandparents, is potentially a traumatic experience that may play a role in ASD development. A ‘premigration theory of autism’ is then proposed.