Immersive art focuses on deep audience engagement and experience, enabling immersion in the created environment through multiple inputs from the audience's behaviour, body and mind. The design of immersive space requires the use of new media technology and various types of sensors and other equipment, of which XR (Extended Reality) technology in VR, AR, MR virtual and reality fusion of technology and multimodal emotion recognition for immersive art to provide a technical basis for the experience of self-perception and physiological data projection into the scene of the work to form an interactive, rich in the forms of immersive art and enhance the accuracy of emotion recognition. The immersive works constructed by XR technology, compared with traditional creative media and interactive devices, not only pursues narrative and audience's physical interaction as the goal, but also explores the interactivity of perception. In this essay, we will use the immersive interactive artwork ‘Dream’ as a case study to analyse the artistic expression of immersive interactive video installation, explore how to use multimodal emotion recognition technology to improve the accuracy of emotion recognition and analyse the relationship between sleeping posture, dream and emotion, the narrative design of the dream content image, as well as the interactive behaviour and immersive experience, to explore more dimensions of human-computer interaction under the multimodal emotion recognition, and discuss the future research direction of artistic expression combined with XR technology.