古地理学是一门强数据依赖性学科，古地理重建作为古地理学的核心任务之一，着眼于研究地质历史时期地球表面的地理、生物、气候面貌及其演化规律。随着大数据时代的来临，海量古地理数据的不断积累和计算机技术的高速发展使得标准化、智能化的数字古地理重建成为可能。文章通过介绍国内外与古地理相关的代表数据库及团队，总结其优缺点，提出大数据驱动下的数字古地理重建核心思路：（1） 建立标准化的古地理学知识体系；（2） 建立开放互动、动态更新的古地理数据库，并利用机器阅读技术等拓展数据来源；（3） 建立标准化的古地理学数据质量控制体系；（4） 利用机器学习技术建立各类型古地理重建模型，深度挖掘数据；（5） 以可实时更新的智能数字地图集或多维动画形式输出成果。

Paleogeography is a typical data-reliable subject. Paleogeographic reconstruction focuses on the characteristics of geographic, life, and climate changes on the Earth's surface through geological history. In the new era of big data, the continuous accumulation of massive paleogeographic data and the rapid development of computer science technology make it possible to reconstruct the paleogeographic history using more standard and intelligent tools and software. The current paper reviews the major databases and research groups related to paleogeography, and proposes the following key components of big data-driven paleogeographic reconstruction: (1) A standard paleogeographic knowledge system; (2) An open and interactive paleogeographic data platform with new technologies such as natural-language understanding to expand data sources; (3) Paleogeographic data quality control mechanisms; (4) Various types of paleogeographic reconstruction models contructed with artificial intelligence technology; (5) Visual outputs as time-sliced maps or animations.