Background: Systemic lupus erythematosus (SLE) is chronic autoimmune disease with multiple organ damage and is associated with poor prognosis and high mortality. Identification of universal biomarkers to predict SLE activity is challenging due to the heterogeneity of the disease. This study aimed to identify the indicators that are sensitive and specific to predict activity of SLE. Methods: We retrospectively analyzed 108 patients with SLE. Patients were categorized into SLE with activity and without activity groups on the basis of SLE disease activity index. We analyzed the potential of routine and novel indicators in predicting the SLE activity using receiver operating characteristic curves and multivariate logistic regression. The Spearman method was used to understand the correlation between albumin to fibrinogen ratio (AFR), prognostic nutritional index (PNI), AFR+PNI model and disease activity.Results: SLE with activity group had elevatory C3, ESR, CRP, D-dimer, fibrinogen, CRP to albumin ratio, positive rate of anti-dsDNA and ANUA, lower TBIL, TP, albumin, albumin/globulin, creatinine, HDL-C, hemoglobin, hematocrit, lymphocyte count, AFR, PNI. A further established model based on combination of AFR and PNI (AFR-PNI model) showed prominent value in distinguishing SLE with activity patients from SLE without activity patients. In addition, the sensitivity and specificity of AFR-PNI model +anti-dsDNA combination model were superior to AFR-PNI model. AFR and PNI were risk factors for SLE activity. Moreocer, AFR, PNI and AFR+PNI model correlated with disease activity. Furthermore, AFR, PNI, and AFR-PNI model were associated with fever, pleurisy, pericarditis, renal involvement. Conclusion: These findings suggest that predictive model based on combination of AFR and PNI may be useful markers to identify active SLE in clinical practice.