Financial systems are complex systems which composed by many adaptive interactive agents． During the past 20 years，agent-based computational finance explores the complex operation laws and evolution properties of financial system and expands the traditional classical financial theory through microscopic modeling method of the " bottom-up" ． In the paper，employing the bibliometric method，we generally reviewed the status of publications on agent-based computational finance． And then，based on the individual complex evolutionary perspective，we summarized the development of three important directions i． e． ，the individual learning，the species switching and the complex network and prospected the future research．