This study explored the association between plasma levels of multiple metals and cognitive impairment (CI) in 455 aluminum electrolysis workers from a northern Chinese plant, divided into CI (256) and control (199) groups by MoCA scores. Using inductively coupled plasma mass spectrometry, 11 metals were measured, with analyses via conditional logistic regression,generalized linear models (GLM), Bayesian kernel machine regression (BKMR), and age stratification (40 years). Plasma aluminum (Al), lead (Pb), lithium (Li), manganese, cobalt, and copper were significantly higher in the CI group (all P < 0.05), while zinc showed no difference. Single-element analysis found Al, Pb, and Li negatively correlated with MoCA total and subscores (e.g., visuospatial function; P < 0.05), and zinc positively correlated with attention ( β = 1.10, P < 0.05). BKMR confirmed metal mixtures above the 25th percentile reduced MoCA scores ( β = -0.875, 95 % CI: -1.379 to -0.371), with Al, Pb, and Li as key contributors (PIP > 0.6). Subgroup analysis showed Al primarily affected those < 40, while Pb had greater impact in those > 40. Findings indicate elevated Al, Pb, and Li associate with higher CI risk, metal mixtures synergistically exacerbate impairment, and age modifies these effects, aiding occupational cognitive impairment prevention.