Objectives As maps increasingly participate in social construction, cultural communication, and knowledge production, narrative cartography has emerged as a foundational expressive practice. However, existing studies have focused mainly on print-based or early digital maps, leaving the theoretical foundations of narrative maps under intelligent digital media underexplored. With the rapid development of intelligent and digital technologies, including new media forms (e.g., animation and games), extended reality（XR), and the metaverse, narrative maps have undergone profound transformations in their media form, narrative logic, and meaning production. In this context, it becomes necessary to clarify the fundamental theoretical issues of intelligent digital narrative maps (IDNM). This study aims to address three core questions: How IDNM should be conceptually defined and distinguished from print-based forms. How their narrative structures are organized under conditions of interactivity and computation. And through what mechanisms narrative meaning is generated in human⁃machine interaction.

Methods A qualitative, theory-oriented approach was employed, based on a comprehensive survey of representative IDNM practices. Through inductive reasoning, key operational and expressive features distinguishing IDNM from print-based maps were distilled. On this basis, a conceptual typological framework was developed, structured around core dimensions. The study further examined how dynamic interactions among cartographers, users, and machines co-produce spatiotemporal frameworks and narrative meaning.

Results IDNM are defined as human⁃machine assemblages integrating hardware, digital systems, and users, operating under algorithmic logic, and generating meaning through embodied interaction and multimodal symbol construction. Three core characteristics of IDNM are identified: interactivity, self-organization, and assemblage. Building on these characteristics, a typological cube model is established, enabling systematic classification of IDNM across interactive perspective (from omniscient to first-person experiential), assembly hierarchy (from low-level abstract interaction to high-level embodied interaction), and plot generation (from fixed embedded narratives to emergent narratives). Narrative structures shift from closed, stable forms toward open, dynamic, and generative configurations, while discourse evolves from static visual expression to multimodal, interactive, and procedurally mediated forms. Regarding meaning-generating mechanisms, three key mechanisms are identified: Reconstruction of body-centered spatiotemporal perception, immersive integration of emotion and cognition facilitating flow experiences, and algorithmic constraints guiding narrative emergence. Together, these explain how narrative meaning arises dynamically rather than being fully predetermined by the cartographer.

Conclusions By clarifying the conceptual definition, typology, narrative structure, and meaning-generating mechanisms, we advance a theoretical framework for understanding IDNM. The findings highlight a shift from representation and persuasion toward participation and emergence, repositioning users as active co-producers of narrative meaning. The proposed framework offers a foundation for future empirical research, design practice, and interdisciplinary exploration of IDNM.