A cladistic analysis based on parsimony is undertaken to test hypotheses concerning the monophyly of the ctenid spiders and the relationships among their various clades. The data matrix comprised a total of 98 species representing 16 families scored for 146 characters, with all but six taken from various morphological systems; the remaining are behavioral attributes.

Ctenidae is shown to be polyphyletic as currently delimited, but the most severely questioned taxa (i.e., Acanthoctenus, Cupiennius, and Ancylometes) are indeed members of this family. The so-called ctenid eye pattern is shown to be a synapomorphy uniting a more restricted set of taxa. This clade is recognized as Ctenidae sensu stricto and it comprises at least five major lineages: Acantheinae, Acanthocteninae, Calocteninae, Cteninae, and Viridasiinae. The latter taxa are hypothesized to be the sister group of all other ctenids. In turn, a clade formed by Miturgidae s.s. and Zoridae s.s. is proposed as the sister group of Ctenidae. The cladistic analysis also refutes the monophyly of Ctenus, Anahita, Enoploctenus, Celaetycheus, and Leptoctenus.

The following taxonomic changes are proposed: Anahita isaloensis Ono, 1993 is transferred to Vulsor Simon, 1888; Diallomus Simon, 1897, currently placed in Zoridae, is transferred to Ctenidae; Tunabo Chamberlin, 1916 is synonymized with Neoctenus Simon, 1897; Neoctenus is removed from Zoridae and transferred to Trechaleidae; Xenoctenus Mello-Leitão, 1938 is removed from the synonymy of Tunabo and considered incertae sedis, within Ctenoidea.