Multiple myeloma (MM) is a prevalent hematologic malignancy characterized by abnormal proliferation of cloned plasma cells. Given the aggressive nature and drug resistance of MM cells, identification of novel genes could provide valuable insights for treatment. In this study we performed machine learning in the RNA microarray data of purified myeloma plasma cell samples from five independent MM cohorts with 957 MM patients, and identified O-GlcNAcylation transferase (OGT) and cell division cycle 27 (CDC27) as the key prognostic genes for MM. We demonstrated a close link between OGT and CDC27 in MM cells by knockdown of OGT with siOGT, pharmacological inhibition of O-GlcNAcylation with OSMI-1 and pharmacological accumulation of O-GlcNAcylation with Thiamet G. Using mass spectrometry and immunoprecipitation, we identified the O-GlcNAcylated CDC27 protein as a key target protein that may be directly downregulated by OSMI-1 in MM.1S cells. We further revealed that O-GlcNAcylation maintained CDC27 protein stability by blocking the autophagy-lysosome pathway (ALP). Moreover, we demonstrated the enhanced antitumor efficacy of combined OSMI-1 and bortezomib (BTZ) treatment in MM cells both in vivo and in vitro. Thus, this study identifies a novel function of O-GlcNAcylation-related ALP in regulating CDC27 protein stability and a potential therapeutic strategy for treating MM.