Multiple myeloma is a plasma cell cancer with poor survival, characterized by the clonal expansion of multiple myeloma cells (MMC), primarily in the bone marrow. Novel compounds are currently tested in this disease, but partial or minor patients' responses are observed for most compounds used as a single agent. The design of predictors for drug efficacy could be most useful to better understand basic mechanisms targeted by these drugs and design clinical trials. In the current study, we report the building of a DNA methylation score (DM score) predicting the efficacy of decitabine, an inhibitor of DNA methyltransferase (DNMT), targeting methylation-regulated gene expression. DM score was built by identifying 47 genes regulated by decitabine in human myeloma cell lines and the expression of which in primary MMCs of previously untreated patients is predictive for overall survival. A high DM score predicts patients' poor survival, and, of major interest, high sensitivity of primary MMCs or human myeloma cell lines to decitabine in vitro. Thus, DM score could be useful to design novel treatments with DMNT inhibitor in multiple myeloma and has highlighted 47 genes, the gene products of which could be important for multiple myeloma disease development.