Social risks posed by infrastructure projects can seriously threaten the successful implementation of the project and the social stability of the city more broadly. There is thus a need to better understand how social risks are originated and developed so that decision makers can better manage them. In this study, nine social risk factors and twelve stakeholders associated with these social risks are identified based on a literature review and 43 infrastructure cases. Social networks among these stakeholders in social risks are formalized using social network analysis (SNA). To reflect the evolution of social risk evolution more accurately, the behaviors of the general public in five different behavior states (appeal, pussyfoot, probably protest, protest, withdraw) are simulated by means of agent-based modeling (ABM), with the developed model being applied to a real-world case for demonstration purposes. Two scenarios are evaluated in terms of the communication effects and the risk governance effects in the evolution of social risk. This SNA-based ABM model can be a powerful tool for testing the response effectiveness of different risk mitigation strategies with respect to the impacts of the social risk development process, as well as for supporting social risk management decision making in infrastructure projects.