Multiple myeloma is the second largest malignant tumor of the blood system. Proteasome inhibitors (PIs) currently are effective drugs for some myeloma patients, but their prognosis varies. We extracted the transcriptome expression data and clinical information of myeloma patients from MMRF CoMMpass database, and used the Random Survival Forest Variable Hunting (RSF-VH) algorithm to select 6 highly prognosis-related genes and to develop a 6-genes scoring model, by which the risk score predicted were significantly associated with the progress-free survival (PFS,  ). The median PFS of the high-risk group is 21 months, while it is 29 months in the low-risk group. The scoring model was further validated in the testing cohort. Furthermore, Analysis revealed that the risk score performed better in predicting the multiple myeloma patients’ prognosis than the existed staging system, including R-ISS. The risk score is independent with the most existed clinical risk indicators, and the prognostic effectiveness of 6-genes scoring model is homogenous in patients with different clinical observations. Further bioinformatic analysis revealed that the risk score is not only significantly associated with multiple myeloma-related pathways, including immune response, but also with the infiltration of many kinds of immune cells that associated with clinical malignancy. Collectively, the model we developed using one lincRNA and five mRNAs is a robust and effective indicator for myeloma patients’ prognosis undergoing proteasome inhibitors therapy.