Objective: Dyslipidemia is a major risk factor for atherosclerosis development. Evidence suggests atherosclerotic plaque accumulation can begin in childhood and lead to adverse cardiovascular events into adulthood. The use of metabolomics to identify metabolites specific to dyslipidemia may provide insights into the pathogenesis of cardiovascular-related outcomes. Thus, our aim was to identify metabolite networks associated with dyslipidemia in childhood. Methods: We used cross-sectional data from 326 children (median age of 5.2 years) from Gen3G, a prospective pre-birth cohort. We evaluated 1,038 plasma metabolites (798 annotated and 240 unannotated). Lipid measures included total cholesterol, non-HDL, LDL, triglyceride, and HDL. Using the AHA and the American Academy of Pediatrics guidelines, we identified normal and borderline/high classifications for each lipid. We applied weighted-correlation network analysis to find highly correlated metabolite networks. Spearman’s partial correlations were applied to assess the associations of lipids with metabolite networks, adjusting for age and sex, with false discovery rate correction. Results: We identified a green module of 120 metabolites, mainly comprised of lipids, that showed positive correlations with total cholesterol, non-HDL, and LDL classified as borderline/high vs normal (ρ adjusted = 0.28 to 0.40) and as continuous measures (ρ adjusted = 0.32 to 0.63), while HDL (borderline/high vs normal) was inversely associated ( Figure ). In this module, sphingomyelin (d18:2/16:0, d18:1/16:1)* and sphingomyelin (d18:2/23:0, d18:1/23:1, d17:1/24:1)* appeared to be driving the associations. Also, the blue and yellow modules showed inverse correlations (mainly driven by 10-nonadecenoate (19:1n9) and 3-hydroxyhexanoate, respectively), while the brown module showed positive correlations with triglyceride (driven by 1-palmitoyl-GPC (16:0)). Conclusions: A unique network of metabolites was associated with dyslipidemia in childhood.