Fossils are invaluable information resources for understanding the deep-time Earth history. Over hundreds of years, a huge amount of paleontological data recording those information has been published. With the rapid development of computer, database and internet technologies during the past 30-40 years, those data have been gathered into various paleontological databases under different goals. The databases all have distinguished system structure, data organization method and service objects. In the present study we reviewed the development of major paleontological databases around the world, including their history, architecture, data characteristics and data volume. Their data organization methods, key online functions, data sharing mechanisms as well as the quality control technique of taxonomic data have also been compared and evaluated. Moreover, several cutting-edge data-driven paleontological research have been introduced. Based on the experience of their data application routine, a concept of establishing a harmonized paleontological big data platform containing data compilation, standardization, sharing, analysis and application, was proposed. It can serve as an example in the Deep-time Digital Earth (DDE) Big Science Program for the construction of multi-disciplinary geosciences big data platform.