Background: Dyslipidemia is associated with and often precedes the onset of chronic kidney disease (CKD). However, standard lipid panels fail to describe the full spectrum of blood lipidome. A longitudinal profiling of the full spectrum of blood lipidome associated with the risk of CKD is still lacking in any racial/ethnic group, including American Indians, a population that suffered a disproportionately high burden of CKD. Objective: To identify novel lipids and lipidomic signatures associated with risk of CKD in American Indians, independent of traditional risk factors. Methods: We included 1,910 American Indians who attended two examinations (~5.5 years apart) and were free of CKD and cardiovascular disease (CVD) at baseline. CKD was defined based on the estimated glomerular filtration rate (eGFR) calculated by the CKD-EPI formula. Fasting plasma levels of 1,542 lipids in 3,916 samples at two-time points were repeatedly measured by untargeted LC-MS. Machine learning (elastic net) was used to identify lipidomic predictors of CKD beyond traditional risk factors (age, sex, center, BMI, hypertension, diabetes, urinary albumin/creatinine ratio, HDL, triglyceride, and use of lipid-lowering medications). Mixed-effect linear models were used to examine the relationship between changes (baseline to 5-year follow-up) in lipidome and change in kidney function (eGFR), adjusting for traditional risk factors mentioned above, baseline lipid, and baseline eGFR. Multivariate analysis was conducted to identify discriminatory lipidomic signatures that can differentiate between high- and low-risk individuals. Results: Of 1,910 participants free of CKD and CVD (17.8% diabetic) at baseline, 55 (2.9%) developed incident CKD by the end of a 5-year follow-up. Elastic net identified 106 (24 known) lipids, largely glycerophospholipids (GPs), sphingomyelins (SMs), triacylglycerols (TAGs), fatty acids (FAs), and acylcarnitines (ACs), whose altered baseline levels were associated with the risk of CKD, independent of traditional risk factors. A composite lipid score composed of the 24 known lipids improved CKD risk prediction beyond traditional risk factors (AUROC improved from 0.944 to 0.963, P = 0.006). Longitudinal changes in 182 lipids (75 known) lipids, largely GPs, FAs, SMs, TAGs, ACs, and diacylglycerols (DAGs), explained up to 4.8% of the variation of the eGFR change. Multivariate analysis identified distinct lipidomic signatures that separated high- from low-risk participants. Conclusion: We identified novel molecular lipids that significantly predict CKD onset and progression among American Indians beyond traditional risk factors. Our findings shed light on the mechanisms by which dyslipidemia affects CKD and provide instrumental data for biomarker identification, risk prediction and stratification, prognosis, and potential therapeutics.