Keloids are a skin fibrosis disease characterized by troublesome symptoms, a varying degree of recurrence and inevitable side effects from treatments. Thus, identifying their drug targets is necessary. A 2-sample Mendelian randomization analysis was conducted using proteins from the intersection of the deCODE database and "The sure variable. The outcome variable was based on recently published GWAS of keloids. Summary data-based Mendelian randomization and colocalization analysis was employed to distinguish pleiotropy from linkage. Candidate targets underwent drug target analysis. The primary findings were validated through single-cell RNAsequencing data, Western Blot and immunofluorescence staining on keloids. Seven proteins were identified as potential drug targets for keloids. Among these proteins, Hedgehog-interacting protein, neurotrimin [NTM], KLKB1, and CRIPTO showed positive correlations with keloids, while PLXNC1, SCG3 and PDGFD exhibited negative correlations. Combined with the single-cell RNA-sequencing data, NTM, PLXNC1, and PDGFD were found highly expressed in the fibroblasts. NTM showed a significant increase in keloids as compared to normal scars. In accordance with the analysis, higher levels of protein expression of NTM in keloids compared to normal skin was observed. The identified proteins may be appealing drug targets for keloids treatment with a special emphasis on NTM.