The paper addresses the issue of optimizing ore feedstock base to operate iron and steel works. It is shown that fluctuating ore feedstock base assessed on the basis of drum cold mechanical strength index TI can result in falling short in achieving the optimal economic values of agglomeration and blast-furnace process. A method of combined use of the stress model of solid inclusion in matrix and the failure model based on the Rittinger's law is suggested to account for effects of varying feedstock agglomerates on agglomeration and on efficiency of the blast-furnace process. The metallurgic agglomerate failure model also accounts for combined use of the stress model developed for solid inclusions in matrix. The algorithmic issues dealing with complex models integration are addressed. The calculation results are provided to estimate the change in the blast furnace production efficiency when agglomerates made of various ore materials are used. The comparison of agglomerates having different strength expressed through TI index is made.