Methane is one of the most important greenhouse gases and plays an essential role in atmospheric chemistry. The largest single source of methane is natural wetlands, which have been suggested to contribute significantly to the interannual variability of global methane emissions. Methanogens are one of the main functional microbial groups mediating methane cycles of natural wetlands. Biogenic methane is produced by methanogenic archaea or methanogens as the final step in anaerobic degradation of organic matter. It is evident that the studies on the diversity of methanogens can assist with revealing microbial-mediated methane cycles and the temporal-spatial heterogeneity of methane emission from natural wetlands. Traditional methods based on laboratory culture techniques have been proven inadequate to describe the vast microbial diversity, because those methods miss more than 99% of the organisms. Molecular methods independent of culture techniques has vastly improved the potential to describe microbial diversity. Both phylogenetic (16S rRNA gene) and functional genes can be used as molecular markers in the study of methanogens. Methyl-coenzyme M reductase (MCR) is a vital enzyme in methane production, and the mcrA gene coding for a subunit of MCR has been employed as a specific marker for the detection and differentiation of methanogen communities. Sequence-based mcrA phylogeny is consistent with the 16S rRNA-based phylogeny. Thus, the mcrA gene is a favorable functional gene and widely used to detect methanogens in soils of natural wetlands. To understand the key factors affecting methane production, the variation in community structure of methanogenic archaea in the soils from different depths was investigated using PCR-RFLP technique and sequence analysis in the Phragmites australis marsh soil in the Min River estuary. One hundred clones were randomly chosen from each clone library to examine by Colony PCR. A total of 591 positive clones were determinately contained mcrA. The PCR-RFLP analysis showed that 591 positive clones were classified into 37 different operational taxonomic units (OTUs). 37 clones were chosen to sequence and compared with sequences in the GenBank database, the most similarity is between 91%-99%. RFLP analysis and phylogenetic analysis showed that the communities of methanogens in the P.australis marsh soil could be subdivided into three groups which are Methanobacteriales, Methanomirobiales and Methanosarcinales. The distribution of methanogens community in the different depths showed different feature. The dominant group in the soil surface layer (0-10 cm) was Methanoregula, accounting for 76% of total methanogens. The main groups present in the 10-20 cm soil layer were Methanolinea and Methanoregula which account for 23% and 29% of total methanogens, respectively. In the 20-30 cm soil layers, the dominant group was Methanolinea which account for 66% of total methanogens. Shannon index (H' ) and Simpson diversity index (D) showed that the diversity of methanogens in the 10-20 cm soil layer was higher than that in other two layers. In addition, 26 OTUs in the 37 sequenced clones were belongs to the uncultured methanogens, which suggested that there exist a large number of uncultured methanogens in the P. australis marsh soil in the Min River estuary.