Abstract: 【Objective】The purpose of the study was to compare the three molecular identification methods of arbuscular mycorrhizal(AM) fungi based on rDNA sequence analysis, and to provide a reference for improving the accuracy of molecular identification of AM fungi at the species level. 【Method】Morphological identification of AM fungi(numbered A6, B3 and GY3-1) isolated from the rhizosphere soil of common crops such as citrus and sugarcane in Guangxi was carried out. The 18S rDNA, 28S rDNA and sequences containing ITS1, 5.8S rDNA and ITS2(abbreviated as ITS1+5.8S+ITS2) of AM fungi were amplified by nested PCR respectively. The sequences obtained by sequencing were respectively aligned with GenBank data and a phylogenetic tree was constructed. The accuracy of the three molecular identification methods based on rDNA sequence analysis was then compared with the morphological identification results. 【Result】Morphological charaters of A6, B3 and GY3-1 were identical with Claroideoglomus etunicatum, Septoglomus viscosum and Funneliformis mosseae respectively. Based on 18S rDNA sequence, strain A6 was identified as Claroideoglomus, strain GY3-1 as F. moessae, and strain B3 as S. viscosum. Based on 28S rDNA sequence, strain A6 was identified as C. etunicatum, strain B3 as S. viscosum, and strain GY3-1 as F. moessae. Based on the ITS1+5.8S+ITS2 sequence, strain A6 was identified as C. etunicatum, strain B3 as S. viscosum, and strain GY3-1 as F. moessae. Combining the results of the three molecular identifications, the identification results of A6, B3 and GY3-1 strains were C. etunicatum, S. viscosum and F.mosseae respectively. 【Conclusion】All three methods can be used for the molecular identification of AM fungi at the species level. The identification methods mainly based on 18S rDNA and 28S rDNA are relatively simple and fast, and are more suitable for the identification of AM fungi at the genus level. The identification method mainly based on ITS has more complicated steps, but the accuracy of identification to the species level is higher. The analysis results of the three methods can complement each other to a certain extent and enhance the accuracy of molecular identification result of AM fungi.