Diabetic kidney disease (DKD) poses a significant health challenge for individuals with diabetes. At its initial stages, DKD often presents asymptomatically, and the standard for non-invasive diagnosis, the albumin-creatinine ratio (ACR), employs discrete categorizations (normal, microalbuminuria, macroalbuminuria) with limitations in sensitivity and specificity across diverse population cohorts. Single biomarker reliance further restricts the predictive value in clinical settings. Given the escalating prevalence of diabetes, our study uses proteomic technologies to identify novel urinary proteins as supplementary DKD biomarkers. A total of 158 T1D subjects provided urine samples, with 28 (15 DKD; 13 non-DKD) used in the discovery stage and 131 (45 DKD; 40 pDKD; 46 non-DKD) used in the confirmation. We identified eight proteins (A1BG, AMBP, AZGP1, BTD, RBP4, ORM2, GM2A, and PGCP), all of which demonstrated excellent area-under-the-curve (AUC) values (0.959 to 0.995) in distinguishing DKD from non-DKD. Furthermore, this multi-marker panel successfully segregated the most ambiguous group (microalbuminuria) into three distinct clusters, with 80% of subjects aligning either as DKD or non-DKD. The remaining 20% exhibited continued uncertainty. Overall, the use of these candidate urinary proteins allowed for the better classification of DKD and offered potential for significant improvements in the early identification of DKD in T1D populations.