Abstract: The world is undergoing momentous changes unseen in a century, with a significant increase in uncertainties facing policymaking. As a reinforcement learning mechanism to cope with uncertainties, policy experimentation promotes policy innovation on the local scale and tests the effect of new programs to enhance the understanding of policy issues and the effectiveness of programs, reduce decision-making risks, and improve the applicability and validity of public policies. Due to the differences in uncertainties, policy experimentation has differentiated application scenarios and functional orientations. Existing studies have focused on the purposes and types of policy experimentation, analyzed the organizational model, cognitive logic and functional orientation of policy experimentation, and constructed diverse theoretical models. The most influential models include democratic experimentalism, experimentalist governance, experimental learning and adaptive governance. This paper presents the knowledge accumulation of existing studies through literature review, which helps to deepen the understanding of the effect of policy experimentation in coping with uncertainties at the theoretical level, and has also application value for optimizing the mechanism design of policy experimentation at the practical level.