AIM: To explore the causal links among circulating inflammatory proteins (CIPs) and the varying severities of diabetic retinopathy (DR). METHODS: This research utilized a two sample Mendelian randomization (MR) approach to explore the causal relationships between 91 CIPs and various severities of DR: background DR (BDR) or non-proliferative DR (NPDR), and proliferative DR (PDR). Single-nucleotide polymorphisms (SNPs) related to the 91 CIPs as exposure factors were identified. These SNPs were selected from an extensive genome-wide association study (GWAS) analyzing large genomic datasets. Genetic variation data of various DR phenotypes provided by the FinnGen collaboration were utilized as outcomes. Inverse-variance weighting (IVW) was used as the main MR analysis. Robustness of study results was evaluated through a series of sensitivity analyses, employing the MR-pleiotropy-test and mendelian randomization pleiotropy residual sum and outlier (MR-PRESSO) to confirm the absence of pleiotropy. RESULTS: In a bidirectional MR analysis, we uncovered a complex relationship between CIPs and DR. Elevated levels of tumor necrosis factor ligand superfamily member 14 (TNFSF14), latency associated peptide transforming growth factors beta-1 (LAP-TGF-beta1), interleukin-10 (IL-10), and vascular endothelial growth factor A (VEGF-A) were associated with a reduced risk of NPDR. Conversely, elevated levels of fibroblast growth factor 23 (FGF-23) were associated with an increased risk of NPDR. Concentrations of adenosine deaminase (ADA), matrix metalloproteinase-10 (MMP-10), eotaxin, and IL-10 showed elevated levels and were linked to a reduced risk of NPDR. On the other hand, the levels of oncostatin-M, beta-nerve growth factor (β-NGF), and interleukin-7 (IL-7) were elevated and associated with an increased risk of SNPDR. Elevated levels of ADA, MMP-10, and macrophage colony-stimulating factor 1 (CSF1) were linked to a lower likelihood of PDR. Conversely, elevated levels of Caspase 8 and glial cell line-derived neurotrophic factor (GDNF) were associated with an increased risk of PDR. In reverse MR analysis, DR affected the expression of these factors. CONCLUSION: Our research demonstrates evidence supporting a potential causal link between key inflammatory factors and the risk and prognosis of various DR phenotypes. These findings emphasize the regulation of inflammatory factors responses as a strategic approach for preventing and managing DR. Altogether, our results validate the pathogenic role of inflammatory factors dysregulation in DR and support the rationale for exploring immunotherapeutic targets further.