Early identification of young children with an autism spectrum disorder (ASD) can lead to earlier entry into intervention programs that support improved developmental outcomes. The purpose of the present study was to examine identification and diagnostic patterns of children with ASD who live in a large metropolitan area. One hundred fifteen 8-year-old children diagnosed with ASD were identified from a population-based surveillance system at the Centers for Disease Control and Prevention. Primary variables of interest included earliest age of evaluation and earliest age of diagnosis identified from surveillance records, type of initial ASD diagnosis, evaluation sources that documented first ASD diagnosis, characteristics of professionals assigning first ASD diagnosis, and diagnostic tools used to aid the diagnostic process. We found that children with ASD identified by the surveillance system were initially evaluated at a mean of 48 months but were not diagnosed with ASD until a mean age of 61 months. There were no differences in timing of diagnosis based on sex or racial/ethnic classification, although degree of impairment associated with ASD predicted mean age at first evaluation and mean age at first ASD diagnosis. Most children were identified at non-school sources, such as hospitals and clinics; 24% of the sample did not receive a documented ASD diagnosis until entering school. Most practitioners (70%) did not use a diagnostic instrument when assigning the first ASD diagnosis. Implications for early identification of ASD are discussed.