ABSTRACTBackground and AimsOral ulcer (OU) is a complex issue with limited effective treatments. This study uses multi‐omics data through summary Mendelian randomization (SMR) and colocalization analysis to identify specific gene associations with OU, aiming to find new therapeutic targets, repurpose existing drugs, and develop new treatment options.MethodsOur study consists of two phases: first, extracting data from Genome‐Wide Association Studies and using blood mQTL, eQTL, and pQTL data as exposure factors, then integrating these with OU gene data through SMR analysis. Then, we validate the results with UK Biobank data and perform colocalization analysis to confirm shared genetic variants.ResultsGenetically predicted levels of four circulating proteins are associated with OU. Under strong supportive evidence from mQTL, eQTL, and pQTL, genetically predicted levels of NFKB1 are negatively correlated with the risk of OU. With moderate supportive evidence from mQTL and pQTL, genetically predicted levels of FAIM3 are negatively correlated with the risk of OU. Meanwhile, under low supportive evidence from eQTL and pQTL, higher genetically predicted levels of JUND and lower levels of IL12β are associated with a higher risk of OU.ConclusionSMR approach employed in this study has pinpointed several proteins with tangible associations to the risk of OU. NFKB1, FAIM3, JUND, and IL12β stand out as promising therapeutic targets for OU, beckoning further exploration and research.