Integrated flood risk management involves a large portfolio of options for mitigating risks that include hard (engineered methods such as dams and flood-ways) and soft structural (nature based such as wetland protection, upper watershed restoration and rain gardens), non-structural, and recovery responses. An example of non-structural responses are flood warnings, emergency services supported by individual and collective actions and the use of mitigation and resilience measures. Agent-based model (ABM) approaches allow for the representation of social interaction in human and natural domains and for the investigation of the emergence of adaptive, collective responses to changing environmental management practices. Agent Based Modelling is applied in a variety of contexts; from physical modelling, socio-ecological modelling and social behavior modelling, to more complex modelling such as coupled human and natural systems or environmental modelling. Recent studies show that climate change and urbanization are projected to increase flood risks. Flood risk management is becoming increasingly important. It is important to take into account human and environmental interaction for flood risk analysis in order to avoid misestimating of the future flood risk. Such issue requires a successful operation of complex, socio-technical interaction systems for effective flood risk management. The aim of this study was to develop an Agent Based Model (ABM) and integrate it with hydraulic model for simulating the human-flood interaction to evaluate the effectiveness of flood risk communication systems and social vulnerability, and determine the critical factors that govern the effectiveness of the evacuation process due to extreme flood events, taking into account the road network, the information dissemination among evacuees, local information centres and the lead time of the early warning system. The influence of flood risk management policies on flood risk was also analysed.