Anthurium is a strictly Neotropical genus of Araceae ranging from southern Mexico to northern Argentina, including approximately 900 accepted species names. Despite its immense diversity, its ecological importance in Neotropical forests, and a long history of botanical collection, cultivation, and taxonomical research, Anthurium has been only cursorily sampled in previous molecular phytogenies. This study aims to test the monophyly of Anthurium, to understand the evolutionary history of the genus, and to elucidate relationships among its species using maximum parsimony, maximum likelihood and Bayesian analyses on a combined chloroplast (trnG intron, tmH-psbA and tmC-ycf6 intergenic spacers) and nuclear (CHS first intron and partial flanking coding regions) DNA sequence dataset for 102 Anthurium species and closely related outgroups. Despite some limitations (ca. 11% species richness coverage, and lack of nuclear sequences for outgroups), results indicate that Anthurium is a strongly supported monophyletic genus and that at least 18 major supported clades are recognizable within it, most of them easily characterized morphologically and/or geographically. This study also suggests that the current sectional classification of Anthurium does not accurately represent its evolutionary history since most of the major clades recovered in these analyses do not correspond with the current circumscriptions of infrageneric groups. Despite using the most variable gene regions available, low sequence divergence was found among Anthurium species, relatively short branches characterize the core of the Anthurium clade, and resolution is still lacking in the deeper nodes of the phylogeny, a pattern consistent with a rapid, and probably recent, radiation of species.