Libinia, a genus of spider crabs, encompasses 10 valid species, three from the eastern Pacific and seven from the western Atlantic. Previous phylogenetic hypotheses based on molecular and larval morphological data suggested an uncertain position of the genus Libinia within the family Epialtidae. Therefore, we investigated the evolutionary relationships among the species of Libinia, and also suggested a position for this genus within the family. Phylogenetic trees were obtained from concatenated mtDNA sequences of 16S and COI fragments. Molecular dating methods were used to estimate the relative timing of origin and diversification within Libinia. Eighteen new primers were designed, and PCR conditions were standardised to amplify the COI fragments of older or poorly preserved pinned specimens, generating ~550 base pair sequences. Phylogenetic hypotheses supported the monophyly of the genus. Species that occur in the eastern Pacific and the ones found in the western Atlantic do not form monophyletic clades related to this geographical separation. In the eastern Pacific, there is low phylogenetic proximity between Libinia mexicana and L. setosa; and since we did not sample L. peruana, we inferred that it should cluster with L. setosa or L. mexicana due to morphological similarity. Libinia spinosa is separate from the other species, L. mexicana was a sister group of L. cavirostris, and L. ferreirae was closely related to L. dubia and L. rhomboidea, which occur in the Gulf of Mexico and in the Caribbean, respectively. In the Gulf of Mexico, L. dubia and L. emarginata are closely related. Although L. erinacea was not sampled, we suggest, from previous studies, that it should group with L. dubia due to morphological similarities. The phylogenetic hypothesis suggests the monophyly of Libinia, with its origins at the late Eocene. Finally, the recovered clades based on two molecular markers agreed with a parallel, linked morphological study that is still in progress.