The present invention belongs to the electric power system hydroelectric generation dispatching field, and discloses a reservoir group joint optimized dispatching feasible decision space identification method which enables the calculation amount and the storage amount of a conventional method to be reduced, realizes a considerable dimensionality reduction effect, and has a very good application prospect in the reservoir group joint dispatching field. The technical scheme of the method comprises firstly determining demands, constructing a reservoir group optimized dispatching model, and determining an objective function and constraint conditions; then, adopting the specific knowledge rules to identify the feasible domains of the reservoirs at a single stage, two stages and multiple stages, and realizing the effective reduction of the search space; finally, adopting the current existing reservoir group optimized dispatching methods, such as a dynamic programming method, a particle swarm optimization algorithm, etc., to solve. The method of the present invention can pre-process the constraint conditions automatically, identifies the feasible decision space of the reservoir group dispatching problem scientifically, enables the calculation amount and the storage amount of the conventional method to be reduced and a dimensionality curse problem to be relieved greatly, and provides a reasonable and feasible method for the efficient solution of the reservoir group joint optimized dispatching problem.