The article deals with the issues of modeling socio-economic systems at the regional level. A promising direction in this area is the use of an agent-based approach based on taking into account the features and a certain autonomy of the behavior of interacting agents. Moreover, one of their key properties is the ability to adapt to changing conditions of functioning. Proceeding from this, in order to solve the problems of increasing the efficiency of developing strategic decisions, it is necessary to develop tools to support management decisions, which will be based on the use of an adaptive simulation model of the region. To do this, the article provides a brief overview of the types of agents depending on the methods of processing and using the information they perceive, an algorithm for the functioning of the agent within the framework of the simulation model is presented. It is shown that the information received by the agent can be of different nature. Separate characteristics of any objects, phenomena form a database. The collection of such information forms a precedent. A set of precedents form a rule base. The analysis showed that the adaptability of the agent's behavior is becoming an essential condition for the implementation of his goals. It should be noted that the agent acts in conditions of incomplete information, and his actions are largely described by the theory of bounded rationality. The structure of perceived information is shown and an approach to reflecting the adaptive properties of an agent through the learning subsystem is proposed. Its elements and implementation options are highlighted. From the point of view of reflecting the adaptive properties of the agent, obtaining new information leads to the replenishment of the database and the base of use cases. At the same time, within the framework of analytical activities, new information is compared with the expectations formed on the basis of previously available data and reflected in the agent's rule base. If the information received is contrary to the existing rules, then this becomes the starting point for their revision. Thus, the adaptation of the agent within the framework of the developed simulation model of the region is based on the adjustment of the rule base.