The primary objective of this study was to investigate the effect of dietary fiber on methanogenic diversity and community composition in the hindgut of indigenous Chinese Lantang gilts to explain the unexpected findings reported earlier that Lantang gilts fed low-fiber diet (LFD) produced more methane than those fed high-fiber diet (HFD). In total, 12 Lantang gilts (58.7 ± 0.37 kg) were randomly divided into two dietary groups (six replicates (pigs) per group) and fed either LFD (NDF = 201.46 g/kg) or HFD (NDF = 329.70 g/kg). Wheat bran was the main source of fiber for the LFD, whereas ground rice hull (mixture of rice hull and rice bran) was used for the HFD. Results showed that the methanogens in the hindgut of Lantang gilts belonged to four known species (Methanobrevibacter ruminantium, Methanobrevibacter wolinii, Methanosphaera stadtmanae and Methanobrevibacter smithii), with about 89% of the methanogens belonging to the genus Methanobrevibacter. The 16S ribosomal RNA (rRNA) gene copies of Methanobrevibacter were more than three times higher (P < 0.05) for gilts fed LFD (3.31 × 10 9 copies/g dry matter (DM)) than gilts fed HFD (1.02 × 10 9 copies/g DM). No difference (P > 0.05) was observed in 16S rRNA gene copies of Fibrobacter succinogenes between the two dietary groups, and 18S rRNA gene copies of anaerobic fungi in gilts fed LFD were lower than (P < 0.05) those fed HFD. To better explain the effect of different fiber source on the methanogen community, a follow-up in vitro fermentation using a factorial design comprised of two inocula (prepared from hindgut content of gilts fed two diets differing in their dietary fiber) × four substrates (LFD, HFD, wheat bran, ground rice hull) was conducted. Results of the in vitro fermentation confirmed that the predominant methanogens belonged to the genus of Methanobrevibacter, and about 23% methanogens was found to be distantly related (90%) to Thermogymnomonas acidicola. In vitro fermentation also seems to suggest that fiber source did change the methanogens community. Although the density of Methanobrevibacter species was positively correlated with CH 4 production in both in vivo (P < 0.01, r = 0.737) and in vitro trials (P < 0.05, r = 0.854), which could partly explain the higher methane production from gilts fed LFD compared with those in the HFD group. Further investigation is needed to explain how the rice hull affected the methanogens and inhibited CH 4 emission from gilts fed HFD.