<p>Circadian rhythms and depressive disorder (DD) are closely connected. Disrupted circadian rhythms can lead to or exacerbate symptoms of depression. From a transcriptomic perspective, this study aimed to explore the diagnostic potential of circadian rhythm-related genes (CRGs) as biomarkers for DD. Total RNA was extracted from whole-blood samples from 15 adolescent patients with DD and 15 age-matched controls collected under fasting conditions in the morning. Following RNA sequencing and quality control, a post hoc power analysis was performed to assess sample adequacy. The differential expression analysis revealed differentially expressed genes (DEGs) that intersected with a circadian rhythm-related gene set. A PPI network was constructed to identify hub genes. Diagnostic performance was evaluated in the training set using ROC curves and tenfold cross-validation and then validated in two independent datasets (GSE76826 and GSE32280). Biomarkers with a consistent AUC &gt; 0.7 and expression trends were incorporated into a nomogram, which was assessed using calibration curves, the Hosmer–Lemeshow test, and decision curve analysis. GSEA, CIBERSORT immune cell infiltration analysis, and a predicted lncRNA–miRNA–mRNA network were used to explore the functions and mechanisms of the biomarkers. RT–qPCR was performed to validate gene expression in an independent cohort. CCL23 and VNN1 were identified as key circadian rhythm-related DD biomarkers and were consistently upregulated in samples from multiple datasets. They showed high diagnostic accuracy in the discovery cohort (AUC: 0.916 for CCL23 and 0.987 for VNN1) and maintained an AUC &gt; 0.7 in the two independent validation cohorts. A nomogram integrating both biomarkers exhibited excellent predictive performance and clinical utility. Enrichment analysis revealed that DEGs were associated with pathways including “cardiac muscle contraction”, “ribosome” and “neuroactive ligand-receptor interaction”. CCL23 was linked to the “proteasome” and “long-term potentiation”, and VNN1 was linked to the “spliceosome” and “oxidative phosphorylation”. Both biomarkers were coenriched in “benzodiazepine receptor activity”, implicating the GABAergic system. Immune profiling revealed altered infiltration of 10 immune cell types in DD patients, with a positive correlation between the expression of CCL23 and the number of naive B cells (r = 0.5; P = 0.0046). Single-cell RNA sequencing indicated low but detectable expression across major neuronal and glial cell types. A predicted ceRNA network suggested that regulators such as AC000120.1 and AC079781.5 may target VNN1 via miRNAs (e.g., hsa-miR-1185-5p). RT‒qPCR confirmed the significant upregulation of the expression of both biomarkers in DD patients compared with controls (<i>P</i> &lt; 0.05). Two CRGs (CCL23 and VNN1) were identified as biomarkers for DD and exhibited excellent diagnostic performance. These findings provide promising prospects for future DD research.</p>