Roller compaction is a pharmaceutical unit operation typically scaled-up and controlled through trial and error to produce compacted powders with specific qualities. The goal of this research was to develop a methodology to predict final roller compact properties based on raw material properties and process conditions for use in scale-up and control. Uniaxial compacts were created with an Instron/Sintech universal testing machine to simulate roller compaction ribbons. The viscoelastic nature of microcrystalline cellulose (MCC) and mixtures of acetaminophen (APAP) and MCC were successfully represented by viscoelastic solid behavior through stress relaxation experiments and Maxwell modeling. Viscoelastic parameters from Maxwell models and average powder particle size were used to model the viscoelastic contact area between two deforming particles in the system. The viscoelastic contact area was coupled with process parameters from uniaxial compaction in a statistical least squares regression models to predict experimentally determined final compact density with a correlation coefficient of R² = 0.90 at 90% confidence. This model is greatly improved over a model without contact area of R² = 0.66. Contact area was shown not to improve statistical models for density of roller compacted ribbons. This study shows that viscoelastic properties in least squares regression models in conjunction with process parameters is a viable method for control of a uniaxial compaction process. Forces, such as friction and particle size distribution, not included in this study are thought to have a greater effect on roller compaction than the viscoelastic properties.