The economic decisions concerning fire safety of built property may involve alternative solutions (alternatives) of fire protection measures, comparison of buildings with different fire safety levels, choice among construction products with different properties of performance in fire. The need to choose among alternatives may be faced by various interested parties: fire safety regulators (authorities on national level), insurers, architects (building designers), manufacturers of fire protection measures, property owners and buyers. The problem of choice will often involve the need to consider simultaneously several characteristics (attributes) of alternatives, and those related to fire safety will be accompanied by economic and non-economic ones. Such a choice can be formalised as a problem of multi-attribute selection (MAS), a filed known also as multi-criteria decision making (MCDM). The present paper aims to formulate and solve several problems of multi-attribute selection by taking into account attributes related to fire safety. The problems are considered in an attempt to facilitate decision-making on three levels: the level of a property buyer/renter, the level of a property owner, and the level of an architect (building designer). It is shown that a different level of decision making requires to apply attributes of different nature. The well-developed field of fire safety offers a range of quantitative and qualitative indicators describing fire performance of building materials, construction and fire protection products as well as entire buildings. These indicators are introduced, with relative ease, into MAS problems. The attributes of economic nature fit naturally in the MAS problems involving fire safety. They can be introduced passively and accompany fire-related attributes as well as other non-economic attributes. However, economic aspects of MAS problems can be used in a more subtle way, namely, by introducing elements of formal expression of risk posed by fires. Monetary losses caused by fires are among the elements of a risk profile. An expected value of these losses can be used as an MAS attribute.http://dx.doi.org/10.5755/j01.ee.22.3.516