With the growing adoption of Immersive Virtual Environments (IVEs) in fields such as cultural education, engineering guidance, and architectural review, improving both the efficiency and quality of IVE creation has become increasingly important. The prevailing workflow in industry typically follows a sequential pattern: editing on a desktop (PC), reviewing in VR, and then returning to the PC for further adjustments. This repeated switching between devices not only increases time and physical effort, but also introduces contextual reconstruction and cognitive discontinuity, ultimately limiting overall productivity.

　　Although recent research and tools have explored creating IVEs directly within VR, limitations in input precision and control stability make it difficult for VR alone to handle the full range of editing tasks. Integrating the embodied spatial perception afforded by VR with the precise symbolic control provided by PC platforms therefore presents a promising direction for improving IVE creation workflows. Against this background, this study addresses the following core question: how can a VR–PC cross-device collaborative system be designed to optimize the existing IVE creation process?

　　Based on a comprehensive literature review and expert interviews, this study conceptualizes the creation of an IVE as a process centered on fundamental “transformation operations” and their combinations. A within-subject experimental study was also conducted to systematically compare the performance of VR and PC in representative transformation tasks. The results indicate that PC is better suited for parameterized input and constrained axial transformations requiring stable control, whereas VR performs better in tasks that rely on embodied spatial perception, such as spatial positioning and multi-object alignment. These findings reveal structural differences in the cognitive affordances of the two platforms.

　　Building on these results, and informed by Distributed Cognition Theory, Joint Activity Theory, and CSCW (Computer-Supported Cooperative Work) perspectives, this study proposes four key design principles for VR–PC cross-device collaboration: cognitive complementarity, shared state and real-time synchronization, operational mapping equivalence, and low coordination cost. Guided by these principles, a prototype VR–PC collaborative system for IVE creation was designed and implemented.

　　At last, this study used the virtual exhibition construction as the experimental scenario. A between-group comparative study was conducted, combining quantitative and qualitative analyses. Evaluation metrics included task completion time, completion rate, rework frequency, and device-switching cost. The results demonstrate that, compared with the traditional single-user sequential workflow, the VR–PC collaborative approach achieves better overall person-time efficiency, greater task stability, and improved control over rework. The findings also support the effectiveness of the proposed design principles.