The usefulness of randomly amplified polymorphic DNA method (RAPD) to identify each species of genus Erysipelothrix and for epidemiological analysis of this genus was studied. Eighty-one strains and 18 random primers were tested. Among the tested primers, the primers NK51 (GGTGGTGGTATC) and NK6 (CCCGCGCCCC) produced noticeable results. The primer NK51 revealed four species-specific RAPD patterns. Of the 66 strains of E. rhusiopathiae, 64 had the same unique band of 884 bp. Of the 12 strains of E. tonsillarum, 11 produced a 1,265-bp band. In addition, two strains, previously thought to be E. rhusiopathiae, produced the 1,265-bp band, suggesting that they had been misclassified. One strain of E. tonsillarum produced the 884-bp band, suggesting that it too was E. rhusiopathiae. The E. rhusiopathiae strain of serovar 13 produced a 650-bp band, and the strain of serovar 18 produced a clear 420-bp band as well as three weak bands of 1,265, 918, and 444 bp. The primer NK6 revealed 14 RAPD patterns that were not serovar specific. However, different patterns were produced among strains of the same serovar showing that the RAPD method is able to identify the genetic variations of strains of this genus and can rapidly and easily differentiate strains of the same serovar. Based on these results, we concluded that the RAPD method with primers NK51 and NK6 is a rapid and reliable method to identify the species of this genus; we also concluded that this method might be a useful tool for the epidemiological analysis of the Erysipelothrix species.