Despite likely being the most diverse group within the Tricladida, the systematics of land planarians (Geoplanidae) has received minor attention. The most species‐rich ingroup, the subfamily Geoplaninae, is restricted to the Neotropics. The systematics of Geoplaninae remains uncertain. Unique features supporting the genera are scanty; moreover, parts of the known species have been poorly described, making comparative studies difficult. Likewise the evolutionary relationships among land planarians remain insufficiently understood. In the present study, a phylogenetic hypothesis for selected taxa of Geoplaninae based on the molecular data is presented and discussed in the light of morphological features. Our phylogenetic inference is based on the fragments of three nuclear regions (18S, 28S rDNA and EF‐1α) and a mitochondrial marker (cytochrome oxidase I) for which we considered three optimality criteria (parsimony, maximum likelihood and Bayesian inference). Although our data provide little support for most basal nodes, our phylogenetic trees show a number of well‐supported clades, unveiling morphologically homogeneous groups. According to these results, we propose to separate Geoplana into Barreirana (formerly considered a subgenus), Cratera gen. n., Imbira gen. n., Matuxia gen. n., Obama gen. n. and Paraba gen. n., emend the diagnoses of Barreirana, Geoplana, Notogynaphallia, Pasipha and Xerapoa and review the classification of the species within these genera. For Geoplana goetschi sensu Marcus, (1951), a new name is proposed.