The diagnostic process of autism spectrum disorders in Hispanic children from bilingual home environments may produce results in false positives or false negatives and thus delay in diagnosis and intervention. A gap in the literature exists in the assessment of Hispanic children from bilingual home environments. The present study is based on the enactive mind theory that posits social communication is constructed through social interactions which may be affected by language acquisition in the bilingual home. A quasi-experimental static-group design was used in the study. Archival data from 75 consecutively referred Hispanic children were examined. Archival video clips of the administration of the Autism Diagnostic Observation Schedule (ADOS) were rated using the PDD-Behavior Inventory (PDD-BI) to determine if the language pragmatic score predicted group membership of children with and without autism spectrum disorders. The enter and forward methods of logistic regression were employed to ascertain if an additional pragmatic language difference score increased the accuracy of prediction of the groups and which scores, the PDD-BI, the ADOS, or the Autism Diagnostic Interview-Revised (ADI-R) increased the prediction of the group membership. Results indicated that parent responses about behavior on the ADI-R resulted in a higher percentage of correctly classified cases. Improved diagnosis will result in earlier interventions that can improve language, behavior, and independent functioning for Hispanic children with autism disorders. Future research of these measures in other samples of English language learners will increase diagnostic capability and improve outcomes in other populations.