In the face of uncertainty and dynamic banking environment and the impact of a large number of external and internal factors on commercial banks, banks are interested in choosing the most appropriate banking strategies, which allowed achieving their goals at the least cost and potential losses. The traditionally used technique of multivariate cluster analysis enables the efficient cluster modeling only in statics but not in dynamics. Moreover, in the conditions of the unstable economy of the developing countries like Kazakhstan this technique gives incorrect results. To address these shortcomings, the authors have developed a new technique of dynamic multi-dimensional cluster analysis, which are free of these shortcomings. In the study, authors investigated the strategic behavior of banks in Kazakhstan for 2009-2015, using this technique. The study helps to shed some light on the formation of the banks development strategies in Kazakhstan. The technique of dynamic multi-dimensional cluster analysis proposed by the authors is especially useful to professionals in the field of banking strategies or other persons, who is interesting in the development prospects of banking sector.