Purposeis development of the methods to predict indices of iron-ore deposits relying upon the improvement of available techniques as well as formulation of new geometrization procedures and identification of the most adequate decision-making way to assess geological data as the basis for geometrization and prediction. Methods are to develop a self-organizing prediction algorithm based upon combination of the available techniques and formulation of new mathematical methods; consider various means to assess them in the context of iron-ore deposit; and select the most efficient one. Use of geostatistical methods makes it possible to evaluate and process output geological information. The methods help assess mineral reserves of a mining enterprise. Findings. Dependencies of magnetite ore content upon geological factors have been derived in the context of an open pit of PIVDGZK JSC. The deposit has been geometrized; predictive mining and geometric model of the deposit site has been deve-loped. Factors have been determined influencing the distribution nature of the indices. Graphs to arrange grade indices of the deposit have been constructed. The graphs have helped predict their placement within the deposit. Originality. A method to predict mining and geological indices of iron-ore deposit has been developed relaying upon a self-organizing algorithm. Correlation between grade indices of minerals and different geological factors has been determined making it possible to describe spatial distribution of grade indices of the deposit. Practical implications. Geometrization methods for iron-ore deposits have been formulated. The methods help schedule mining operations accurately while improving their efficiency. The developed predictive self-organizing algorithm is the flexible tool used for various mining and geological conditions to provide scheduling and assessing of different mining methods. The self-organizing as well as geostatic evaluation techniques is quite a promising research tendency.