Purpose <p>Shared mental models (SMMs) reflect the extent to which team members hold a similar understanding of roles and team processes. In trauma resuscitation, higher SMM alignment is considered a prerequisite for coordinated interprofessional teamwork. This study compared the effects of checklist-based versus algorithm-based in situ simulation training on teamwork-related SMMs in trauma room teams.</p> Methods <p>In this multicenter, cluster-randomized pre–post study, 29 interprofessional trauma teams (186 participants) from six hospitals received either checklist-based training (<i>n</i> = 15 teams) incorporating the Trauma Room Manual checklists or algorithm-based training (<i>n</i> = 14 teams), both grounded in ATLS principles. Teams completed an SMM questionnaire adapted to the trauma setting, based on a previously published instrument, assessing similarity in (1) task responsibility and (2) team communication. Team-level similarity scores were calculated pre- and post-training. Training effects were analyzed using mixed repeated-measures ANOVA.</p> Results <p>Across both training formats, teamwork-related shared mental models increased significantly from pre- to post-training in total similarity, task responsibility, and communication. No significant between-group differences or time × group interactions were observed. The interaction for task responsibility approached significance (<i>p</i> =.06; η²<i>p</i> =.12).</p> Conclusion <p>A single-day in situ simulation training improved teamwork-related shared mental models in interprofessional trauma teams. Checklist- and algorithm-based formats produced similar short-term gains. Future studies should evaluate long-term retention, implementation fidelity, and whether improved SMM alignment translates into measurable team performance and patient-safety outcomes.</p>