The article presents the development of information-logical system that allows the recognition of blast furnace deviation kinds from the normal mode. The effect of melting course parameters of blast furnace at the risk of deviations from normal operation has been investigated. A functional model of information system, consisting of 94 decomposed blocks, as well as a three-component extensible, scalable software architecture, consisting of a unit of work with databases, mathematical libraries and a graphical user interface have been developed. The resulting software allows detecting operatively the presence of deviations from the normal course of melting, determining its type on the actual data of the operation furnace and taking actions to eliminate these deviations. The functionality of the software and the results of the tests for managing blast furnace at JSC “Magnitogorsk iron and steel works” were described.