Abstract: Five species (Phoebe bournei, Phoebe zhennan, Phoebe sheareri, Machilus oreophila, Machilus pauhoi) were chosen to extract timber DNA by using DNeasy Plant Mini Kit(Qiagen), CTAB method and SDS-CTAB method. Based on the effect of DNA extraction, this study explores the method of DNA extraction from Phoebe wood, and several kinds of wood were identified by DNA barcode. The results show that good quality DNA from fresh wood could be extracted by all the 3 methods. Among them, the DNA extracted from the kit has a small amount of protein residue. The DNA extracted by CTAB method and SDS-CTAB method has overall good quality and a small amount of RNA exists, but the difference in extraction effect is not significant. After the wood was dried at 30 ℃, 70 ℃ and 103 ℃, DNA extraction was carried out. It was found that the DNA extraction amount after drying was significantly decreased. Only traces of wood DNA was extracted after drying at 103 ℃, but still meets the needs of PCR amplification. In summary, CTAB method is the best method for DNA extraction from Phoebe wood. After that, four DNA section of matK, trnL-trnF, trnL-intron and rpoB from 5 species were amplified by PCR, and PCR product were sequenced and aligned. The results showed that 4 DNA barcodes do not completely distinguish the 5 species.There are 5 polymorphism sites on trnL-intron sequence between 5 species. Based on the difference, two different genera could be distinguished. Besides, M. oreophila and M. pauhoi could be distinguished while the other 3 could not be. There are 3 polymorphism sites on matK sequence, based on which M. oreophila and M. pauhoi could be distinguished while the other 3 could not be. There are 2 polymorphism sites on both of rpoB and trnL-trnF sequence which could distinguish just between 2 different genera but not species. From the phylogenetic tree obtained by the NJ method based on trnL-intron sequence, five species can be correctly clustered as a reference for distinguishing these tree species.