Effects of male-biased dispersal on inbreeding avoidance were investigated in a semi-natural population of Myodes (formerly Clethrionomys) rufocanus using a large outdoor enclosure (3 ha). Parentage of 918 voles weaned from 215 litters and relatedness of mates were analysed using microsatellite loci, and dispersal distances were obtained from mark–recapture live-trapping data. Natal and breeding male-biased dispersal was observed. There remained, however, chances that incestuous mating could occur, because not all males dispersed from their natal site, and 51 matings occurred between relatives (relatedness r > 0). The number of weaned juveniles from inbred litters was significantly smaller than that from non-inbred litters. Fourteen incestuous matings occurred between close relatives (r ≥ 0.25), most of which were those between non-littermate maternal half siblings (four cases) and those between paternal half siblings (seven cases). When comparing the observed frequencies to the expected ones generated by combining every oestrous female with a male randomly chosen from her surroundings, the observed values for inbreeding of r ≥ 0.25 were significantly smaller than the expectations, while no difference was observed for inbreeding of 0 < r < 0.25. These results suggest that male-biased dispersal is partly effective to avoid incestuous mating, but it does not provide complete separation of male and female close relatives. Additional mechanisms such as kin discrimination based on familiarity may work in inbreeding avoidance of the vole.