BACKGROUND AIMS: Predicting autologous peripheral blood stem cell (PBSC) collection yield before leukapheresis is important for optimizing PBSC mobilization and autologous stem cell transplantation (ASCT) for treating hematological malignancies. Although guidelines for plerixafor usage based on peripheral blood CD34+ (PB-CD34+) cell count are available, their predictive performance in the real world remains unclear. METHODS: This study retrospectively analyzed 55 mobilization procedures for patients with non-Hodgkin lymphoma or multiple myeloma and developed a novel quantitative prediction model for CD34+ cell collection yield that incorporated four clinical parameters available the day before leukapheresis; namely, PB-CD34+ cell count the day before apheresis (day -1 PB-CD34+), number of prior chemotherapy regimens, disease status at apheresis and mobilization protocol. RESULTS: The effects of PB-CD34+ cell counts on CD34+ cell collection yield varied widely per patient characteristics, and plerixafor usage was recommended in patients with poorly controlled disease or those with a history of heavy pre-treatments even with abundant day -1 PB-CD34+ cell count. This model suggested a more proactive use of plerixafor than that recommended by the guidelines for patients with poor pre-collection condition or those with a higher target number of CD34+ cells. Further, the authors analyzed the clinical outcomes of ASCT and found that plerixafor use for stem cell mobilization did not affect short- or long-term outcomes after ASCT. CONCLUSIONS: Although external validations are necessary, the results can be beneficial for establishing more effective and safer mobilization strategies.