Dietary (poly)phenols and methylxanthines may help lower the risk of developing neurodegenerative diseases. However, the relationship between (poly)phenols and methylxanthines and age-related cognitive impairment has not been fully explored, in particular (poly)phenol-rich diets using biomarkers. A cross-sectional analysis was conducted on 202 participants aged 60 - 80 diagnosed with mild cognitive impairment (MCI) or two or more cardiometabolic disorders (CMD). The associations between metabolic signature/metabolites, cognition, and cardiometabolic risk were assessed using linear models, adjusted for covariates and multiple testing (FDR<0.05). The metabolic signature of (poly)phenol-rich diet score (PPS) was linked with Quality of Episodic Memory (QEM), sustained attention, attentional fluctuation and intensity in the subgroups with CMD or MCI (FDR-adjusted P<.05). In plasma, 8 (poly)phenol metabolites presented significant inverse associations with MCI diagnosis, with 4-O-caffeoylquinic acid showing the strongest association (odds ratio: 0.84 (95% CI: 0.72, 0.98)). In cerebrospinal fluid (CSF), 6 metabolites were significantly associated with 4 cognitive outcomes, including QEM (FDR-adjusted P<.05). As for cardiometabolic outcomes, 5 plasma metabolites were inversely associated with Framingham risk score, while (−)-epicatechin in CSF showed negative associations with LDL-C and TC (FDR-adjusted P<.05). Plasma methylxanthine showed favorable associations with cardiometabolic markers, including paraxanthine and HDL-C (stdBeta: 0.17 (0.15, 0.30)), and 3-methylxanthine and HbA1c (stdBeta: −0.25 (−0.43, −0.07)). The PPS signature and (poly)phenol/methylxanthine metabolites from plasma and CSF are significantly associated with favorable cognitive and cardiometabolic outcomes, highlighting the potential role of a (poly)phenol-rich diet, particularly phenolic acids, in promoting cognitive and cardiometabolic health.