The paper considers an approach to the construction of decision support systems based on the use of the subject area semiotic models andmethods of natural language processing. The peculiarity of the proposed approach lies in the fact that the unstructured linguistic information of the Internet is used as the knowledge base of the semiotic model. The structuring of this information is carried out based on a subjective qualitative semiotic model of the situation, built by an expert, and the use of natural language processing methods for texts relevant to the subject area, which are obtained from the Internet. The functional structures of decision support systems in the tasks of monitoring the situation state, making decisions, and supporting the modeling results interpretation in conditions of uncertainty are given. Qualitative assessments of the possibility of implementing decision support systems based on a subjective semiotic model and natural language processing methods are given.Conclusions based on the results of experimentswith a software model of the system for the decision support problem are presented, confirming the effectiveness of this approach.