Facial attractiveness, which impacts every aspect of people's lives, has garnered attention from academics and the general public. Corresponding to the development of artificial intelligence technology, it’s possible to construct computer models for predicting facial attractiveness. Consequently, numerous facial attractiveness databases have been created. The facial features extracted by these databases can be summarized into three categories: formal aesthetics, evolutionary aesthetics and data-driven features; the algorithm of the prediction models can be classified into symbolism and connectionism. According to the face features and modeling method, the aesthetic and algorithmic logic of facial attractiveness database have been reviewed and summarized, moreover, the driving forces behind the development of these databases has been identified. Future research should focus on establishing standardized dynamic databases. Furthermore, it is also important to explore the essence of facial beauty from an interdisciplinary perspective.