In this study, we aimed to evaluate the usability of some genetic markers which were previously reported the capability to identify Fusarium isolates based on barcording with nucleotide sequences, and to clear up the questionable points of application for actual isolates. We constructed a local database containing 46 reference sequences of Fusarium strains already-identified, and sequenced six genetic regions of 19 food-borne Fusarium isolates. And then, the nucleotide sequence homologies of each genetic region were calculated pairwisely between an isolate and a reference strain. The 18S rDNA, 5.8S rDNA, 28S rDNA and ITS1 sequences leaded to the accurate identification of only three to 13 isolates, respectively, because of perfect matches to sequences of more than two species of reference strains, or mis-identification. The lys2 sequences leaded to the accurate identification of several isolates not identified by these four regions. However, other five isolates could not be identified because of non-amplification of lys2 by PCR. The β-tub sequences leaded to the accurate identification of all tested-isolates. Thus, the β-tub is more useful genetic marker for identifying Fusarium isolates in a wide range than other five loci including lys2.