In order to improve the scientific and methodological base of data and measurement systems of testing laboratories within the stage of software products certification, the article suggests an integrated approach to assessing the quality of software. The approach includes structural analysis of information and measurement tasks, implementation of ISO standards when developing information features models and creating databases as well as formalized representation of the vector of features of quality in the course of software testing, based on comparison of current measurements with specified requirements of standards (images). A cognitive model of software product quality was obtained on the basis of the analysis of fuzzy cognitive maps and weakly structured scenarios of interaction between external and internal factors affecting the formation of quality properties of software products. The results of modeling are presented and conclusions are formulated.