We used DNA sequence data from four regions ([1] nrITS; the chloroplast [2] rps16 intron, [3] trnG-S spacer, and [4] trnL-F intron and spacer) to study phylogenetic relationships within tribe Acantheae (Acanthaceae). Our sample includes 18 of 20 recognized genera and 82 of ca. 500 species (plus two Justicieae as out-groups). Results of parsimony and Bayesian analyses were entirely congruent and provided strong support for monophyly of two sub-lineages of Acantheae, referred to here as the ‘one-lipped corolla’ and ‘two-lipped corolla’ lineages, reflecting notable differences in corolla morphology. Subsequent analyses were of the two sublineages separately in order to include all characters (a hypervariable region of trnG-S could not be aligned across the full range of taxa but could be aligned within sublineages). The ‘one-lipped corolla lineage’ comprises six clades of Old World taxa related as follows: [Crossandra (Sclerochiton clade {Cynarospermum [Blepharis (Acanthus clade Acanthopsis)]})]. All presently recognized genera are strongly supported as monophyletic, except that Blepharis dhofarensis is placed with species of Acanthus, with strong support from both parsimony and Bayesian inference (monophyly of Blepharis was rejected by both parsimony and likelihood). Alternate hypotheses based on calyx and androecial morphology regarding Crossandrella and Streptosiphon could not be rejected, but placement of these genera with some species of Crossandra based on pollen was rejected. The ‘two-lipped corolla lineage’ is strongly supported and includes one clade of Old World plants (the Stenandriopsis clade) that is sister to a strongly supported clade that includes all New World Acantheae as follows: [Stenandrium clade (Neriacanthus {Aphelandra lineage})]. The Aphelandra lineage includes the ‘armed’ Aphelandra clade and a polytomy of six unresolved clades: (1) A. squarrosa, (2) Encephalosphaera clade, (3) Geissomeria clade, (4) A. aurantiaca clade, (5) A. pulcherrima clade, (6) Rhombochlamys. In contrast to patterns in the one-lipped lineage, genera in the two-lipped lineage are mostly not monophyletic nor are relationships among them strongly supported by our molecular data or by morphological synapomorphies. We discuss these results in the context of evidence from other sources including macromorphology, palynology, chromosome numbers, and geographic distribution.