Understanding the evolutionary histories of invasive species is critical to adopt appropriate management strategies, but this process can be exceedingly complex to unravel. As illustrated in this study of the worldwide invasion of the woodwasp Sirex noctilio, population genetic analyses using coalescent-based scenario testing together with Bayesian clustering and historical records provide opportunities to address this problem. The pest spread from its native Eurasian range to the Southern Hemisphere in the 1900’s and recently to Northern America, where it poses economic and potentially ecological threats to planted and native Pinus spp. To investigate the origins and pathways of invasion, samples from five continents were analysed using microsatellites and sequence data. Surprisingly, clustering analyses revealed two gene pools. Scenario testing showed independent introductions from an unsampled source, putative source of the second gene pool. The invasion history was much more complex than previously believed, with most of the populations being admixtures resulting from independent introductions from Europe and subsequent spread among the invaded areas. Results also shed light on the micro-evolutionary processes occurring during introductions, and showed that only few specimens gave rise to some of the populations. Analyses of microsatellites using clustering and scenario testing considered against historical data drastically altered our understanding of the invasion history of S. noctilio and will have important implications for the strategies employed to fight its spread. This study illustrates the value of combining clustering and ABC methods in a comprehensive framework to dissect the complex patterns of spread of global invaders.