The issues of organizing multivariate analysis of complex technical systems based on the ideology of data mining DATA Mining as decision support are considered. The final goal of the research is to localize the areas of parameter variation in which a successful solution to the multicriteria optimization problem is expected. The proposed approach not only increases the efficiency of computational experiments, but also makes it possible to identify patterns, i.e. develop templates, as is customary in DATA Mining. The developed methodology involves the use of such DATA Mining tools as processing of low-quality (“dirty”) data, regression analysis methods, clustering methods, and visualization methods. The stages of the methodology are defined and examples of its application for controlled technical objects are shown.