The role of retinol in the prevention of multifactorial chronic diseases remains uncertain, and there is sparse evidence regarding biological actions and pathways implicated in its effects on various outcomes. The aim is to investigate whether serum retinol in an un-supplemented state is associated with low molecular weight circulating metabolites. We performed a metabolomic analysis of 1,282 male smoker participants based on pre-supplementation fasting serum in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study. We examined the association between 947 metabolites measured by ultra-high performance LC-MS/GC-MS and retinol concentration (from HPLC) using linear regression that estimated the difference in metabolite concentrations per unit difference in retinol concentration as standardized β-coefficients and standard errors (SE). We identified 63 metabolites associated with serum retinol below the Bonferroni-corrected P-value (p < 5.3 × 10-5). The strongest signals were for N-acetyltryptophan (β = 0.27; SE = 0.032; p = 9.8 × 10-17), myo-inositol (β = 0.23; SE = 0.032; p = 9.8 × 10-13), and 1-palmitoylglycerophosphoethanolamine (β = 0.22; SE = 0.032; p = 3.2 × 10-12). Several chemical class pathways were strongly associated with retinol, including amino acids (p = 1.6 × 10-10), lipids (p = 3.3 × 10-7), and cofactor/vitamin metabolites (3.3 × 10-7). The strongest sub-pathway association was for inositol metabolism (p = 2.0 × 10-14). Serum retinol concentration is associated with circulating metabolites in various metabolic pathways, particularly lipids, amino acids, and cofactors/vitamins. These interrelationships may have relevance to the biological actions of retinol, including its role in carcinogenesis.