Advanced decision support systems require significant acceleration of decision-making underwhen factors describing the situation are ill-defined and non-metric. As a rule, such conditions arise in the field of politics, culture and in the social sphere. Cognitive modelling technology is applied in these cases. The cognitive models are created by people, experts from different subjects’ fields. The modelling processes take a great deal of time. Furthermore, the result of the modelling has to be verified when the model’s creators cannot get complete information and have to understand the problem very quickly. The factors and their mutual relationships in cognitive models could be verified with Big Data analysis technology. But this approach takes into account only denotational semantics that are based on the mapping of the model on formalised logical constructions, words, objects, schemes. This paper addresses the issue of creating cognitive semantics that take into consideration thinking, feeling andfactors. It is shown that the classical computer or quantum computer cannot ensure cognitive semantics because they are based on discrete representation of data. An optical computing and Optical Semantic approach could be applied. The architecture of the special optical processor is represented