Abstract: Currently, spatiotemporal information, positioning and navigation have become important new infrastructures. Driven by general artificial intelligence, the era of intelligence led by large models has arrived. The increasingly powerful large models will play a more and more important role in the intelligent processing and applications of spatiotemporal information. First, this paper uses large models as the research paradigm and summarizes the current status and progress of large models in intelligent processing of spatiotemporal information in surveying and mapping. In addition, it analyzes the challenges faced by large models in intelligent processing of spatiotemporal information in surveying and mapping. We elaborate three key technologies in spatiotemporal information large models for surveying and mapping, including multi-modal fusion and understanding architecture design, prompt engineering optimization by fine-tuning, and human-in-the-loop guidance for decision-making. Finally, the depth of industry understanding, data security risks, content credibility, and training deployment cost optimization are addressed to predict the development trend of spatiotemporal information large models.