This paper aims to develop an efficient model able to reduce catastrophic consequences and the significant number of victims resulting from fires at construction sites. The paper proposes probabilistic modeling aimed to minimize the probability of failure of a construction site. Methodology: The developed model in this paper consists of modeling fire hazards, the vulnerability of the potential targets, and the risk within construction sites. The optimization algorithm called “differential evolution” is used in order to determine the optimal site layout, which is characterized by having the smallest overall probability of failure. A numerical simulation is performed to delineate an appropriate probability density function of the failure of the site. In addition, a geographic information system (GIS) is used to display the spatial variability of fire risk on a construction site. Findings: The paper provides an efficient model to enhance site layout planning and assign locations for supporting temporary facilities at appropriate positions within a construction site. The model is examined through applying it on a simple case study containing numerous facilities. All these facilities are considered vulnerable targets and some of them are potential fire hazards, with different intensity values. Value: Most of the previous research focuses on travel cost distance in developing site layout planning models. This paper fulfills the development of a valuable model able to generate an optimized construction site layout by minimizing the probability of failure of the whole site. It will assist the decision makers and the risk managers in identifying the riskiest zones on a construction site.