Autobiographical memory is a narrative form grounded in personal experiences. By recalling and organizing past events, individuals can document significant moments, facilitating emotional regulation and stress relief. While extensive research has demonstrated its positive impact on well-being, tools designed to support autobiographical narration remain underdeveloped. Compared to single modality, multimodal forms have been shown to be superior in memory retrieval. However, these methods often struggle to realize self-expression due to the limitations of the user’s creative abilities. To address this gap, we propose an innovative multimodal autobiographical narration system combining large language models (LLMs) and generative AI (GenAI). The system combines audio and visual formats by utilizing an LLM-based agent to guide users in recollection and a GenAI to tackle creative challenges. User studies show the system effectively supports high-quality autobiographical narratives and quickly generates multimodal works aligned with users’ content, providing a powerful tool for self-reflection and creative memory expression.