As aging populations grow, older adults increasingly experience both insomnia and loneliness, significantly impacting their mental and physical health. Traditional interventions primarily rely on in-person social interactions, which may not be accessible to all elderly individuals. This study explores the potential of a multisensory sleep-assistance system that integrates voice companionship, facial expression feedback, and aromatherapy to enhance emotional support and relaxation before sleep. The system simulates intimate social interactions, using familiar voices from loved ones, synchronized aromatic diffusion, and real-time facial expressions to create an immersive and comforting experience. A questionnaire-based survey was conducted with 63 older adults (mean age: 64.2 years) to assess their psychological well-being, sleep quality, and functional preferences. Results revealed a significant correlation between loneliness and insomnia severity (r = 0.53, p < 0.001), highlighting loneliness as a critical factor affecting sleep quality. Functional preference analysis showed that voice companionship received the highest rating, with participants emphasizing the soothing effect of familiar voices. Aromatherapy was also highly rated, whereas facial expression feedback received mixed reactions, particularly among younger older adults who appreciated its interactive warmth. This study demonstrates the potential benefits of multisensory design in sleep assistance, particularly in emotional well-being and relaxation. Future research should incorporate prototype testing and real-world interactions to validate its effectiveness further and explore personalized content and cultural adaptations for improved usability and engagement.