The release of radon gas and progeny from the mining and milling of uranium-bearing ores has long been recognised as a potential radiological health hazard. The standards for exposure to radon and progeny have decreased over time as the understanding of their health risk has improved. In recent years there has been debate on the long-term releases (10,000 years) of radon from uranium mining and milling sites, focusing on abandoned, operational and rehabilitated sites. The primary purpose has been estimates of the radiation exposure of both local and global populations. Although there has been an increasing number of radon release studies over recent years in the USA, Australia, Canada and elsewhere, a systematic evaluation of this work has yet to be published in the international literature. This paper presents a detailed compilation and analysis of Australian studies. In order to quantify radon sources, a review of data on uranium mining and milling wastes in Australia, as they influence radon releases, is presented. An extensive compilation of the available radon release data is then assembled for the various projects, including a comparison to predictions of radon behaviour where available. An analysis of cumulative radon releases is then developed and compared to the UNSCEAR approach. The implications for the various assessments of long-term releases of radon are discussed, including aspects such as the need for ongoing monitoring of rehabilitation at uranium mining and milling sites and life-cycle accounting.