Virtual reality (VR) technology has recently started shaping learning, especially language learning, with the aim of immersing learners into a VR learning environment. However, because of the high system cost of fully immersive VR, desktop VR has been implemented and preferred in educational settings. Based on a constructivist approach, desktop VR has drawn attention to the need for learner autonomy and an authentic VR learning environment. Therefore, this chapter describes empirical research on desktop VR-based learning using a constructivist approach. The research examined university students' interaction and perceptions of learning in this kind of learning environment. Based on the empirical findings gathered from observations and interviews, this chapter has aimed to discuss not only the issues observed both in previous studies and in this chapter, but also additional issues such as scaffolding, self-paced learning, collaboration, and learner differences in order for learning to occur in a well-designed desktop VR learning environment.