As robotics technologies are increasingly integrated into construction sites, cooperation between human workers and collaborative robots needs new strategies to address effective real-time teaming dynamics. Such teaming dynamics depend on strong Team Situation Awareness (TSA)—a collective understanding of environmental changes, task demands, and teammate actions—but traditional monitoring strategies cannot achieve real-time accuracy. This study proposes a multimodal framework for categorizing TSA using unobtrusive, continuous, and real-time psychophysiological data. In a futuristic VR environment featuring dynamic construction hazards, 21 dyads (42 subjects) performed a collaborative pipe installation task. Traditional TSA was measured using the Situation Awareness Global Assessment Technique (SAGAT), while real-time psychophysiological signals were recorded through eye tracking, prefrontal cortex activation, pulse rate, pulse rate variability, and electrodermal activity. Results revealed that teams with high traditional TSA demonstrated a distinct eye-tracking pattern, including increased run count and dwell time on environment-related areas of interest (AOIs) and reduced task-AOIs, suggesting proactive environmental scanning. Additionally, low TSA teams exhibited greater activation in the right prefrontal cortex, indicating heightened attentional control and potential stress response. These findings support strategies for building collaborative robotics systems that respond to humans’ cognitive states, ultimately enhancing safety, efficiency, and coordination in high-risk, data-intensive construction settings.