In order to better understand the use of cognitive architectures within the Human Research and Engineering Directorate of the U.S. Army Research Laboratory, and with the intent of supporting the Land Warrior program, members of the Manned Systems Design Methods Team of the Integration Methods Branch developed an adaptive control of thought-rational (ACT-R) (Anderson & Lebiere, 1998) model of mental manipulation. The goal of the model was to reproduce errors made during mental manipulation, which were similar in type and number to those made by actual soldiers who took a paper-and-pencil test of mental manipulation as part of a cognitive assessment battery. The final model was able to replicate errors made by the soldiers in the mental manipulation test; however, further work and data collection are needed to make the model predictive of the types of errors soldiers could make if they were exposed to specific types of mental manipulation problems. Lessons learned from this initial application of ACT-R are also discussed.