The objective of this study is to model in-vivo propranolol plasma concentration after administrating oral propranolol extended-release (ER) tablets. In-vivo data are typically expensive and scarce. To save time and cost needed to achieve high-quality invivo profile, this work utilizes both in-vitro and in-vivo data. The ensemble of in-vitro and in-vivo data is modeled by stochastic kriging with qualitative factors (SKQ). It treats in-vivo and in-vitro as the two distinct levels of a qualitative factor. By synergistically modeling both types of data, SKQ is able to provide fitted in-vivo profiles whose quality is much higher than those obtained from modeling in-vivo data alone.