This paper proposes an approach for functional knowledge representation based on problemreduction,which represents the organization of problem-solving activities in two levels:reduction andreasoning.The former makes the functional plans for problem-solving while the latter constructs functional units,called handlers,for executing subproblems designated by these plans.This approach emphasizes that therepresentation of domain knowledge should be closely combined with(rather than separated from)its usetherefore provides a set of reasoning-level primitives to construct handlers and formulate the control strate-gies for executing them.As reduction-level primitives,handlers are used to construct handler-associativenetworks,which become the executable representation of problem-reduction graphs,in order to realize theproblem-solving methods suited to domain features.Besides,handlers and their control slots can be usedto focus the attention of knowledge acquisition and reasoning control. This paper proposes an approach for functional knowledge representation based on problemreduction, which represents the organization of problem-solving activities in two levels: reduction and restructuring. The former makes the functional plans for problem-solving while the latter constructs functional units, called handlers, for executing subproblems designated by these plans. This approach emphasizes that therepresentation of domain knowledge should be closely combined with (rather than separated from) its usetherefore provide a set of reasoning-level primitives to construct handlers and formulate the control strategies for carrying them. As reduction-level primitives, handlers are used to construct handler-associative networks, which become the executable representation of problem-reduction graphs, in order to realize the problem-solving methods suited to domain features.Besides, handlers and their control slots can be usedto focus the attention of knowledge acquisition and reasoning control.