Abstract: Building fire spread is a “high-impact, low-frequency” event that may lead to consequences more severe than the initial fire incident, including potential structural collapse. A domino risk assessment method for building fire spread incorporating structural fire design is proposed based on the Bayesian network. The domino effect theory is applied to analyzing the building fire spread process, and an escalation probability calculation method is introduced. Using rooms as nodes and fire spread paths as directed edges, with escalation probabilities as inputs for conditional probability tables, considering the logical relationships between room fires and domino accident levels, a Bayesian network is constructed. Structural vulnerability under domino accident scenarios is evaluated using the Monte Carlo method. Fire spread risk and structural damage risk are assessed through prior and posterior probability analyses. The validity of the method is verified via building fire spread experiments, and its application to constructing the most typical and worst-case fire scenarios, as well as to optimizing design schemes, is demonstrated. The results show that the building fire spread domino risk can be evaluated by the Bayesian network, providing an analytical method and comparative basis for the fire-resistant design of structures.