Abstract: Background　Proteinuria is a core manifestation and a key risk factor for progression in primary glomerular diseases, yet the circadian rhythm characteristics of its excretion remain insufficiently studied.Objective　To apply the generalized additive mixed model (GAMM) to characterize the circadian rhythm of urinary protein excretion in patients with primary glomerular diseases (CKD stages G1-G4) and explore its relationship with pathological type and renal function.Methods　This single-center, cross-sectional exploratory study employed a cross-sectional design. A total of 116 consecutive patients with biopsy-proven primary glomerular diseases, hospitalized at the Nephrology Department of PLA General Hospital from July 2024 to July 2025 were enrolled. Urine samples were collected using a 24-hour intensive, unequal-interval natural voiding protocol, with detailed recording of each voiding time and urine volume. The GAMM was used to analyze the circadian fluctuation of the urinary protein excretion rate. A stepwise modeling approach was applied to verify the existence of the rhythm and to gradually adjust for covariates. The interaction effects of pathological type (IgA nephropathy vs membranous nephropathy) and CKD stage with time were explored. The dilution effect was corrected using inter-voiding interval weighting, and sensitivity analyses were conducted to validate the findings. Results　A total of 116 participants (70 males and 46 females) with a mean age of 45.7 ± 11.5 years were enrolled, and 906 urine samples were analyzed. The GAMM revealed a significant circadian rhythm in urinary protein excretion after full adjustment for confounders (P＜0.001 for the time smooth term), exhibiting an overall 'nocturnal high, diurnal low' pattern. The predicted peak time was 04:54, and the trough was at 13:54. The pathological interaction model indicated differences in rhythm patterns between IgA nephropathy and membranous nephropathy: IgA nephropathy showed a greater rhythm amplitude (peak/trough ratio 1.44 vs 1.36) and a more complex curve shape (effective degrees of freedom, edf = 4.60 vs 3.50). The CKD stage interaction model revealed that as renal function declined, the rhythm amplitude progressively attenuated, tending toward a 'flattened' pattern: significant rhythms with large amplitudes were present in stages G1 and G2 (P＜0.001), marginal significance was observed in stage G3a (P=0.083), while no significant rhythm was detected in stages G3b and G4 (P＞0.05). Sensitivity analyses supported the robustness of these findings.Conclusion　Using the GAMM, this study confirmed a significant circadian rhythm in urinary protein excretion in patients with primary glomerular diseases, characterized by a nocturnal-dominant pattern. The rhythm features were significantly associated with pathological type and CKD stage.