For fire risk assessment, a probability prediction of fire behavior including a small fire to a large one is necessary. However most existing predicting models for fire behavior are considered some factors with dispersion as the decided values, called 'the worst conditions'. In this study, Monte Carlo simulation using a predicting model for fire behavior given two probability parameters, which were the limit heat release rate and the fire spread rate of each unit, was proposed to calculate the probability distributions of design fire. Then some case studies with the effect of fire protection measures were examined and the smoke temperature of the fire room as probability distribution were calculated.