Objective The aim of this study was to identify the variations of canine distemper viruses (CDV) obtained from clinical samples，and provide a theoretical basis for the prevention and control of CDV infection. Methods The segments of hemagglutinin protein (H) gene，fusion protein (F) gene and nucleocapsid protein (N) gene were amplified by RT-PCR. The amplicons were cloned into pGEM-T-easy vector and the recombination clones were sequenced. Subsequently，the nucleotide and amino acid sequences were analyzed. Results The nucleotide lengths of H，F and N genes for the virus were 1863 bp，1653 bp and 2234 bp，respectively. Sequence analysis did not present any nucleotide insertion or deletion. The three sequences determined in this study showed the high amino acid identities corresponding with three Chinese highly pathogenic CDV isolates BJ080127 (H gene)，GN (F gene) and HL (N gene)，which were 97. 3% ，97. 7% and 99. 3% ，respectively. Meanwhile，each of them displayed the nucleotide identity of 94. 4% ，93. 8% and 97. 2% ，when compared with the CDV prototypic strain A75-17. In addition，comparing with the CDV vaccine strain CDV3，they displayed the nucleotide similarities of 88. 5% ，88. 8% and 93. 9% ，respectively. The glycosylation sites analyses revealed a glycosylation site aimed at the 3 ～ 5 amino acids of F gene. The phylogenetic tree displayed that the virus had the closest relationship with the highly pathogenic CDV isolates identified in China.Conclusion In this study，we cloned the H，F and N genes of a CDV determined in canine samples and identified that the virus is a highly pathogenic CDV，which contains a novel glycosylation site in F gene. This study provides evidence in studies on the molecular mutation epidemiology of CDV infection.