Publicly funded mental health services are critical in caring for children with autism spectrum disorder. Accurate identification of psychiatric comorbidity is necessary for effective mental health treatment. Little is known about psychiatric diagnosis for this population in routine mental health care. This study (1) examined correspondence between psychiatric diagnoses reported by mental health clinicians and those derived from a structured diagnostic interview and (2) identified predictors of agreement between clinician-reported and diagnostic interview-derived diagnoses in a sample of 197 children aged 4–14 years with autism spectrum disorder receiving mental health services. Data were drawn from a randomized effectiveness trial conducted in publicly funded mental health services. Non–autism spectrum disorder diagnoses were assessed using an adapted version of the Mini-International Neuropsychiatric Interview, parent version. Cohen’s kappa was calculated to examine agreement between Mini-International Neuropsychiatric Interview, parent version and clinician-reported diagnoses of comorbid conditions. Children met criteria for an average of 2.83 (standard deviation = 1.92) Mini-International Neuropsychiatric Interview, parent version diagnoses. Agreement was poor across all diagnostic categories ( values: 0.06–0.18). Logistic regression identified child gender and clinical characteristics as significant predictors of agreement for specific diagnoses. Results underscore the need for training mental health clinicians in targeted assessment of specific psychiatric disorders and prioritizing treatment development and testing for specific diagnoses to improve care for children with autism spectrum disorder served in publicly funded mental health settings.