Most scientific investigations regarding inflammatory bowel disease (IBD) pathogenesis or therapeutic strategies use dextran sulfate sodium (DSS)-induced models performed on mice. However, differences between human and animal microbiota may confound the data reproducibility from rodent experiments to clinical trials. In this study, the intervention effects of Bifidobacterium longum NSP001 on DSS-induced colitis were investigated using mice colonized with either native or humanised microbiota. Disorders in disease activity index (DAI), morphology and histology of colon tissue, intestinal permeability, and secretion of MPO, TNF-α and IL-6 were ameliorated by daily intake of live B. longum NSP001 cells in both conventional and humanised colitis mice. But the abnormal thymus index, and colonic production of ZO-1 and iNOS were improved only in colitis mice treated with B. longum NSP001 and humanised microbiome. The accumulation of acetic acid and propionic acid in colon microbiome, and the optimization of primary bile acid biosynthesis and glycerophospholipid metabolism pathways in cecum commensals were likely to explain the beneficial effects of B. longum NSP001. These data revealed that intestinal microbiome baseline would possibly affect the manifestation features of interventions by probiotics or dietary components and highlighted the necessity to include humanised microbiome while investigating potential therapeutic strategies based on rodent models.