Aim: To assess whether existing data collection mechanisms can provide accurate and sufficient information about the prevalence of autism in Australia. Methods: Summary data about the number of children aged 0-16 years known to have an autism spectrum disorder (ASD) were gathered from State and Territory health, disability, education sources and autism associations. Summary data were also provided by national sources. Initial contact was made by letter, and follow-up was undertaken by telephone or email. Results: For the years 2003-2004, the estimated prevalence of autism for 6- to 12-year-olds ranged from 9.6 to 40.8/10 000 for the State and Territory data, and from 12.1 to 35.7/10 000 for the national data. There was a similar variation in prevalence estimates for children aged 0-5 and 13-16 years. There was also a variation in prevalence estimates between age groups. Conclusion: Inconsistencies in autism prevalence estimates calculated from existing data sources suggest that further efforts are needed to ensure the collection of reliable information about the prevalence of ASD for national, State and Territory use. Existing data systems need to be improved or additional data systems need to be developed to ensure the collection of reliable information. Reliable and consistent ASD prevalence data would ensure that services are being provided to those who need them and would enhance the opportunities to undertake meaningful population-based research.