Older subjects with acute myeloid leukemia (AML) or those with co-morbid conditions may not withstand conventional conditioning chemotherapy/radiation prior to hematopoietic stem cell (HSC) transplantation. A relatively mild conditioning is therefore being investigated prior to HSC transplantation. This study investigates the potential to describe personalization of conditioning for each subject using nonlinear model predictive control. A nonlinear ordinary differential equation (ODE) mathematical model of the conditioning regimen and HSC engraftment dynamics is presented. Its simulated dynamics are compared with averaged values from a de-identified clinical white blood cell engraftment data. The controller determines the optimal conditioning strategy administered to patients prior to HSC transplantation. A multi-objective controller designs the conditioning regimen to minimize the number of cancerous cells while limiting collateral damage to the patients tissues. The study serves as a preliminary investigation into the capability of mathematical modeling to optimize conditioning prior to HSC transplantation for treatment of acute myeloid leukemia (AML) in humans.