Abstract: In order to reduce the large fair fire risk, explore its fire cause factors, the interpretation structure model is used to analyze the inner link and hierarchical relationship among various factors. Combining cross impact analysis method and Markov chains forecasting model, a quantitative analysis model is built to get the steady- state probability matrix of the fire cause factors. The results show that the Markov probability is more accurate, the model can clarify the key screening and controlling points of large fair fire cause.

Key words: cross- impact analysis, Markov chains forecasting method, interpretative structural modeling, large fair, fire risk analysis