Large quantities of dangerous substances, especially explosive or flammable gases, are processed or stored in hazard installations. A risk-based warning/early-warning system for major hazard installations is very important for the prevention of major accidents. The real-time data-based risk and the risk factors are analyzed in this article, and a fuzzy logic based real-time risk assessment method is proposed. On the basis of fuzzy logic theory, the likelihood of an accident occurrence and the consequence of the accident can be assessed, and the risk value or risk level can be evaluated by utilizing a risk matrix. The method takes advantage of the real-time data acquired from the safety monitoring system so that the change in the risk can be determined as the accident develops. The risk assessment simulation of a vapor cloud explosion (VCE) accident caused by gas leaked from an liquefied petroleum gas tank is performed. It is shown that the risk of a VCE accident varies with the change of the monitored data. © 2008 American Institute of Chemical Engineers Process Saf Prog, 2008