Little is known about the phylogeny of the family Vorticellidae at the generic level because few comprehensive analyses of molecular phylogenetic relationships between members of this group have, so far, been done. As a result, the phylogenetic positions of some genera that were based originally on morphological analyses remain controversial. In the present study, we performed phylogenetic analyses of vorticellids based on the sequence of the small-subunit (SSU) rRNA gene, including one species of the genus Apocarchesium, for which no sequence has previously been reported. Phylogenetic trees were reconstructed with SSU rRNA gene sequences by using four different methods (Bayesian analysis, maximum-likelihood, neighbour-joining and maximum-parsimony) and had a consistent branching pattern. Members of the genera Vorticella (except V. microstoma) and Carchesium formed a clearly defined, well supported clade that was divergent from the clade comprising members of the genera Pseudovorticella and Epicarchesium, suggesting that the differences in the silverline system (transverse vs reticulate) among vorticellids may be the result of genuine evolutionary divergence. Members of the newly established genus Apocarchesium clustered within the family Vorticellidae basal to the clade containing members of the genera Pseudovorticella and Epicarchesium and were distinct from members of the genus Carchesium, supporting the validity of Apocarchesium as a novel genus. Additional phylogenetic analyses of 21 strains representing seven genera from the families Vorticellidae and Zoothamniidae were performed with single datasets (ITS1–5.8S–ITS2, ITS2 alone) and combined datasets (SSU rRNA+ITS1–5.8S–ITS2, SSU rRNA+ITS2) to explore further the phylogenetic relationship between the three morphologically similar genera Carchesium, Epicarchesium and Apocarchesium, using characteristics not included in previous analyses. The phylogenetic trees reconstructed with combined datasets were more robust and therefore more reliable than those based on single datasets and supported the results of trees based on SSU rRNA sequences.