Risk perception plays an important role in risk communication and chemical incident prevention when engaging with different groups of people. Risk perception of lay people emphasizes the individual experience of intuitive judgment and subjective cognition. To study the mechanism of this process on chemical incidents, three causality feedback loops were achieved by analyzing previous studies on this mental process; meanwhile, system dynamics models were constructed. Detailed causal trees of the key state variables (risk perception, risk perspective, risk response, and risk communication) were illustrated and applied to a quantitative simulation using Vensim software. The simulation results reflect the dynamic process of risk perception of lay people, and specific curves of the key state variables were explained to support the rationality of the models. A sensitivity analysis among the state variables and some auxiliary variables were implemented to identify their internal relationships. The proposed methodology quantifies the diversification of psychological activities of risk perception through the simulated data and images, and offers a tool to optimize risk management and the emergency rescue of lay people at risk of a chemical incident. (C) 2017 Elsevier Ltd. All rights reserved.