Abstract: In order to scientifically and rationally determine the fire safety level of subway stations, considering the characteristics of strong subjectivity, ambiguity, and nonlinearity of indicators, a fire safety assessment method for subway stations based on WSR- FPP and cloud model is proposed. Using WSR, and from three dimensions of "physics", "affairs" and "personal management", 6 first- level indicators are obtained. Using triangular fuzzy numbers and fuzzy linear preference programming theory, tolerance parameters are introduced to obtain indicator weights. Using cloud models, and through three cloud models: standard cloud, indicator cloud, and comprehensive cloud, the uncertainty and ambiguity are overcome, and the fire safety level of the subway station is obtained. The result of case study is that the fire safety level of a Zhengzhou metro station is high, which is within the acceptable range. Among them, automatic fire fighting system, manual firefighting system, and personnel evacuation system are the main risk indicators.

Key words: subway station, fire, safety evaluation, Wuli-Shili- Renli, fuzzy preference programming, cloud