Background Systemic lupus erythematosus (SLE) is a complex and incurable autoimmune disease, so several drug remission for SLE symptoms have been developed and used at present. However, treatment varies by patient and disease activity, and existing medications for SLE were far from satisfactory. Novel drug targets to be found for SLE therapy are still needed. Methods Mendelian randomization (MR), an observational study way, was performed to explore potential drug targets for SLE using protein quantitative trait loci (pQTL) from recently published genome-wide association studies (GWAS) of cerebrospinal fluid (CSF) and plasma proteins, which obtained genetic instruments for 154 CSF proteins of 971 participants, and 734 plasma proteins of 23591 participants. Bidirectional Mendelian randomization analysis, colocalization analysis, and phenotype scanning were performed to find key proteins for SLE. In addition, external data verification was implemented to further consolidate the Mendelian randomization findings. Candidate proteins as targets to find drugs and discuss the druggability. Finally, Network pharmacology and molecular docking methods were used to verify the effects of Voclosporin and Cyclosporine on SLE targets. Protein-protein interaction (PPI) and core target analysis of candidate drugs and SLE overlapping targets were performed to identify potential hub targets and interactions. The affinity between drug targets and SLE targets was confirmed by molecular docking. Results In the preliminary analysis, we identified four key proteins as possible drug targets in CSF and plasma proteins, included ICAM-1(P = 4.62E-05, OR = 0.90(0.86, 0.95)), sICAM-1(P = 4.62E-05, OR = 0.49(0.35, 0.69)), FCG2B (P = 7.63E-11, OR = 0.57(0.48, 0.67)), PPP3CA; PPP3R1 (P = 5.47E-07, OR = 0.66(0.57, 0.78)). Among them, ICAM1 was detected in both CSF and plasma proteins. By excluding reverse causality, confounding factors, and linkage disequilibrium (LD), we identified PPP3CA; PPP3R1 as novel drug targets for SLE, including Voclosporin and Cyclosporine. Finally, the Drugbank database shows that novel drugs contain 33 targets for treating SLE. PPI suggested that SIRT1, ACE, PTGS2, and BACE1 were pivotal targets for SLE treatment. In addition, the molecular docking showed that the bioactive molecules of Voclosporin and Cyclosporine had a good affinity with the target of SLE. Conclusions Our integrative analysis suggested that levels of circulating PPP3CA; PPP3R1 had causal effects on SLE risk and served as potential treatment targets. Moreover, this study provides new evidence for Voclosporin as an SLE treatment through Mendelian randomization and Network pharmacology, and warrants further clinical investigation.