An approach to selecting the optimal composition of multicomponent blast-furnace burden based on a set of mathematical models for predicting change-of-state transformations of iron-ore raw materials, as well as for calculating chemical composition and properties of primary and final products of smelting, is presented. The results of predictive-analytical studies of the change in the indices of high-temperature properties of iron-ore materials and formed primary molten slags with respect to the distribution features of burden components along the radius of the throat for burden conditions of one of the blast furnaces are considered. The performed evaluation of the calculated indices of primary melts and their comparison to the technological requirements to iron-ore raw materials is a constituent part of designing an algorithm for solving an optimization problem of selecting the composition of a blast-furnace burden in the expert system under development.