The production of pharmaceutical tablets relies on homogeneous blends of active pharmaceutical ingredients (APIs) with excipients. These are often produced in large tumble blenders which are simple in design. New blending operations are typically studied experimentally, which is time-consuming and expensive for large manufacturing scales. Discrete element method (DEM) simulations of the blending process can enhance the process understanding and reduce development costs. The blending performance of a bin blending process in a compaction line was studied using DEM. The modeled formulation consisted of two active pharmaceutical ingredients (APIs) and an excipient blend. All three components were characterized experimentally and calibrated in the model based on the dynamic angle of repose. The simulation of the entire 25-minute blending operation (20 minutes of pre-blending, 5 minutes of final blending with lubricant) predicted an uneven distribution of the APIs along the container’s rotation axis. An experimental analysis of samples drawn from the process confirmed the predicted trend. The effect of various modifications to the blending process such as the addition of a baffle or changes to the container rotation protocol were tested in simulations. Of those, the most promising strategy to improve the final homogeneity of the blend was to rotate the container 90° around the vertical axis after the pre-blending step. This effectively changed the rotation axis for the final blending step and mitigated the risk of inhomogeneity. This work showcases calibrated DEM simulations of the entire duration of an industrial scale pharmaceutical bin blending process. The results show how DEM simulations can guide the development of industrial processes by ensuring that powder characteristics of the individual components are replicated in the model. This helps to reduce the number of costly experimental trials and the time needed for process development.