Purpose: The purpose of this study is to develop a model for predicting the flow properties of a four-component powder mixture. Method: To build the model, 22 samples were prepared using an extreme vertices mixture design. The flow properties were characterized using rotational shear cell methodology. Two additional blends were tested for external validation to illustrate model applicability. Results: Cohesion was shown to be in a linear relation with unconfined yield strength and a power relation with flow factor. The special cubic model was used to build a mathematical model. Normality test of residuals showed that the regression model was more robust to predict cohesion than to use flow factor. Conclusion: This QbD approach is shown to be useful for predicting flow performance and finding design space during formulation development. (C) 2016 Published by Elsevier B.V.