Phylogenetic relationships, taxonomy and nomenclature issues within East Asian voles Alexandromys were addressed using comprehensive species samples, including all 12 valid species of the genus, and multilocus analysis. We examined the mitochondrial cytochrome b (cytb) gene and three nuclear genes in 36 specimens. Additionally, we examined a data set of 106 specimens using only the cytb gene. Our results did not confirm the aggregation of A. kikuchii, A. montebelli and A. oeconomus into a separate clade, namely the subgenus Pallasiinus. Analysis of incomplete lineage sorting using JML software highlighted both the cases of mitochondrial introgression and incomplete lineage sorting within the genus. Thus, the sister position of A. sachalinensis and A. maximowiczii in mitochondrial trees could be explained by mitochondrial introgression, while the sister position of A. limnophilus and A. fortis in mitochondrial trees could be successfully explained by incomplete lineage sorting. Very short genetic distances, together with an absence of monophyly, of the three species, A. evoronensis, A. mujanensis and A. maximowiczii, is supported by multiple morphological data, which indicates that these three taxa should be one species—A. maximowiczii. Analysis of genetic distances and tree topology revealed that three species of short‐tailed voles—A. middendorffii, A. mongolicus and A. gromovi—are more closely related to each other than to other established species of Alexandromys. The lacustrine vole, A. limnophilus, is closely related to the group of short‐tailed voles. Analysis of the type specimens of limnophilus and flaviventris confirmed that these taxa form one species together with A. l. malygini. Our results suggest that the mountains of western Mongolia are inhabited by a new taxon of short‐tailed voles of the same rank as middendorffii, mongolicus and gromovi—A. m. alpinus ssp. n.