A key challenge in cognitive modeling is capturing how emotional states modulate internal cognitive parameters. While cognitive architectures such as ACT-R (Adaptive Control of Thought-Rational) provide a principled framework for simulating memory and decision-making, their emotional components remain underexplored. This study examines how individual differences in emotional states, particularly anxiety and affective valence, are reflected in core memory-related parameters of ACT-R. Across two experiments using a digits recall task, we introduced emotional variation via affective stimuli and applied a model-fitting procedure to estimate individual values for the mismatch penalty and activation threshold. Results from the second experiment revealed significant correlations between state anxiety and both parameters, suggesting that emotional traits systematically shift memory retrieval dynamics. Our findings offer empirical support for integrating emotion into cognitive architectures without introducing ad hoc modules, and contribute to broader efforts to align the Common Model of Cognition with affective science. This work highlights the potential of inverse modeling as a tool for understanding the emotion–cognition interface and opens new avenues for modeling individual differences in affect-sensitive cognitive systems.