Motivated forgetting is a popular explanation for unwarranted moral self-perceptions. Yet, morality-related motivated forgetting (‘moral forgetting’) lacks conceptual preciseness. Our goal was to advance the conceptual understanding of moral forgetting with a cognitive modeling approach. Specifically, our goal was to test the fundamental conceptual assumptions, namely that moral forgetting represents self-serving systematic information loss (i.e., is immorality- and agency-specific) and that moral forgetting originates from inhibited information retrieval. We used the multinomial storage-and-retrieval model (Riefer & Rouder, 1992) to separate storage and retrieval processes. In an adapted version of Riefer and Rouders’ (1992) incidental paired-associate learning task, we asked participants in a within-subjects design to imagine a sequence of immoral behavior descriptions that referred to the participant as the behavior-performing agent or the behavior-involved patient. After a filler task, participants completed a free-recall-then-cued-recall paradigm. In the recall tasks, participants had to recall capitalized neutral noun pairs that were embedded in the behavioral descriptions. Contrary to our expectation, the model-based results of Experiment 1 (N=119) did not indicate worse retrieval of information associated with the agency-related immoral encoding context. Whereas the modeling approach revealed in fact better retrieval of information associated with the agency-related immoral encoding context, the results, however, supported moral forgetting in the form of worse storage of information associated with the agency-related immoral encoding context. The findings emphasize the value of cognitive modeling as a tool to promote conceptual preciseness in social cognition research. We validate the results of Experiment 1 in a second experiment.