Phenolic compounds, especially flavonoids, are abundant in plant-based foods and are increasingly recognized for their interaction with the human gut microbiota (GM). This study explored the mutual relationship between dietary flavonoids and GM by assessing four flavonoid-rich food matrices (i.e. blonde orange, rooibos tea, dark chocolate, and blueberry) each enriched in specific flavonoid subclasses. After in vitro gastrointestinal digestion, samples underwent 48 h colonic fermentation using the standardized MICODE model inoculated with GM from 3 healthy donors. Microbial and metabolic responses were analyzed via qPCR, 16S rRNA sequencing, and SPME-GC-MS, while flavonoid bioaccessibility and biotransformation were quantified by means of high-resolution mass spectrometry. Blueberry showed the strongest prebiotic potential, increasing the abundance of beneficial taxa (Bifidobacteriaceae, Faecalibacterium prausnitzii, Lactobacillales, and Akkermansia muciniphila) and reducing the abundance of opportunistic groups (Enterobacteriaceae, Desulfovibrio, Clostridium group I). Orange and rooibos exhibited moderate activity, whereas dark chocolate displayed a mixed profile, enhancing some beneficial taxa but also supporting potential pathobionts. Metabolomic analysis revealed diverse low-molecular-weight phenolics resulting from microbial transformation, particularly 3-(3′-hydroxyphenyl)propanoic acid and valerolactones. Functional ecological indicators, including Firmicutes/Bacteroidota ratio, quantitative Prebiotic Index (qPI), and bifidogenic effects, validated blueberry’s superior impact on GM eubiosis. This work confirms the role of flavonoid-rich foods in shaping GM structure and function and underscores the importance of polyphenol type and matrix. MICODE proved effective in elucidating the complex GM–metabolite–food. Future studies integrating meta-transcriptomics and host cell models are warranted to extend these findings to personalized nutrition strategies and microbiome-based dietary interventions.