A knowledge base with a notional model of a subject domain is considered. The notional model consists of a notional structure and a content of notions included in it. The notional structure describes each notion as a result of the generalization or association of other notions. Enumerable and solvable sets of entities define the content of the notions. The notional structure contains the knowledge inference rules and is considered a formal theory of the subject domain that preserves the truth of all the consequences deduced in it. Declarative knowledge is represented by the notions with an enumeration of the entities belonging to them. Procedural knowledge is defined using the knowledge description language, which describes the notions with the solvable sets of the entities and facts and queries to the knowledge base. The fact is a true statement with the logical connectives, parentheses, and elementary expressions: unary predicates of belonging to an entity to the notion, functors for extracting the entities from the composite entities, and relations between the entities. A query is expressed as a fact with free occurrences of an entity variable. The execution of the query creates the notion with the entities for which the fact is satisfiable. The knowledge language provides operations for creating, deleting, and changing the entities and conditional and loop operators. The use of the knowledge description language allows you to increase the expressiveness of presentation and the efficiency of processing “large” knowledge.