In general, because of the unpredictable nature of fire and the various uncertainties related, for example, to material properties at elevated temperature, the reliability of structural fire design can be justifiably questioned. In this project, a typical structural steel design for fire condition is assessed for its reliability. The assessment consists of estimating the probability of failure of structural steel elements exposed to a wide range of fully developed fires. A number of scenarios to account for different passive protection systems and the variability in properties of related parameters are modelled. The main tool of analysis is Monte Carlo simulation using a software named @RISK. The estimated probabilities of failure or reliability indices are measured against acceptable or target values so that definite conclusion with regards to safe or unsafe design can be made. The target probability of failure and the reliability index are also worked out in this project. The overall results show that applying reliability assessment to structural fire design is of great value in pointing out shortcomings in the design and in enhancing the performance assessment of real structures.