Some applications of information granulation and fuzzy-granular computing are considered. An approach to assessing the features (attributes) of malicious activity that characterize various categories of potential threats to the infrastructure and entities of the Smart City is described. This approach combines traditional fuzzy set methods with the capabilities of their granular representation. An analysis of the peculiarities of this approach is carried out, which makes it possible to increase the reliability of assessing the presence of malicious activity in the Smart City infrastructure by clarifying (verifying) fuzzy observable initial data on the features of such activity and expert opinions.