Objective: Neuropsychiatric (NP) manifestations are common in systemic lupus erythematosus (SLE). Because cerebrospinal fluid (CSF) protein patterns may provide insight into NP pathogenesis, this study used a data‑driven approach to determine whether CSF proteomic profiles cluster SLE patients into clinically meaningful subgroups. Methods: CSF samples from 29 female SLE outpatients were analysed using label‑free liquid chromatography–tandem mass spectrometry. Hierarchical clustering of proteomic data grouped patients, and Kruskal‑Wallis and Wilcoxon tests assessed differences in clinical traits among clusters. Weighted Gene Co‑expression Network Analysis (WGCNA) identified modules of proteins, and Pearson correlations linked modules to clinical variables. Results: Three patient clusters were identified. Cluster 1 had the highest frequency of nephritis, depression and cognitive impairment; cluster 2 exhibited alopecia, SSA antibodies and low cognitive impairment; and cluster 3 showed autonomic neuropathy, lupus headache and increased neurofilament light concentration. Six protein modules (M1–M6) were characterised by nervous tissue proteins, lipid‑cycle proteins, macrophage‑derived proteins, plasma proteins, immunoglobulins and intracellular metabolic proteins, respectively. Modules 1–2 correlated with nephritis, depression, longer disease duration and cognitive impairment and were most pronounced in cluster 1. Modules 4–5 showed inverse correlations with cognitive impairment and brain atrophy and were most distinct in cluster 2. Conclusions: Clustering SLE patients by their CSF proteome reveals subgroups that correspond to clinical phenotypes. Differences in proteomic patterns suggest distinct underlying disease processes, indicating that CSF proteomics may help stratify patients and guide personalised management【203749412506946†L75-L111】.