Objective Lupus nephritis (LN) is the main complication of systemic lupus erythematosus (SLE), causing huge financial burden and poor quality of life. Due to the low compliance of renal biopsy, we aim to find a non-invasive biomarker of LN to optimize its predictive, preventive, and personalized medical service or management. Method Herein, we provided a bioinformatic screen combined clinical validation strategy for rapidly mining exosomal miRNAs for LN diagnosis and management. We screened out differentially expressed miRNAs (DEMs) and differentially expressed mRNAs (DEGs) in LN database and performed a miRNA-mRNA integrated analysis to select out reliable changed miRNAs in LN tissues by using R and Cytoscape. Urinary exosomes were collected by ultracentrifugation and analyzed by nano-tracking analysis and western blotting. Detection of aquaporin-2 showed the tubular source of urinary exosomes. Urinary exosomal miRNAs were detected by RT-qPCR and the target of miR-195-5p was verified by using bioinformatic, dual-luciferase, and western blotting. Result 15 miRNAs and their 60 target mRNAs were contained in miRNA-mRNA integrated map. Bioinformatic analysis showed these miRNAs were involved in various cellular biological process. Exosomal miR-195-5p, miR-25-3p, miR-429, and miR-218-5p were verified in a small clinical group (n = 47). Urinary exosomal miR-195-5p, miR-25-3p, and miR-429 were downregulated in patients and miR-195-5p could recognize LN patients from SLE with good sensitivity and specificity, showing good potential in LN disease monitoring and diagnosis. Conclusion We analyzed and obtained a series of differential miRNAs in LN kidney tissues and suggested that urinary exosomal miR-195-5p could serve as a novel biomarker in LN. Further, miR-195-5p-CXCL10 axis could be a therapeutic target of LN.