Abstract: The field of distribution network planning looks forward to achieving intelligent planning, with the vision of achieving fully automated planning with no or minimal human intervention. In the context of digital transformation, the planning of new distribution networks will face three challenges, i. e., diversified graphics, fragmented scenes, and large-scale data. Three key technologies are proposed from the perspective of graph-model-data fusion: the graph-model fusion technology based on electrical drawing recognition and topology intelligent analysis, the model-data fusion technology based on knowledge-driven load/renewable energy deduction analysis and intelligent decision-making, and the graph-data fusion technology based on multi-modal data fusion and multi spatiotemporal data fusion, trying to break down the barriers between theoretical research and digital engineering. Finally, the development of digital planning for future new distribution network is considered and prospected, providing a reference for achieving a large closed-loop model of “machine-centric, human-machine collaboration”.