The invention discloses a method for identifying a hydroelectric generating set model improved type subspace closed loop, and belongs to the technical field of modeling and identification of hydroelectric generating set models. The prediction form simple subspace identification method based on PSO parameter optimization is applied to the hydroelectric generating set model closed loop identification. The method comprises the first step of establishing a hydroturbine speed regulation closed loop system model with output frequency noise; the second step of considering the influence of parameters p and f in the prediction form simple subspace identification method (PARSIM-K) on the identification, and using a PSO algorithm to optimize the parameters of p and f; the third step of using an improved algorithm to identify the closed loop hydroelectric generating set no-load model. The method has the advantages of being fast in algorithm, high in reliability, easy in programming realization and the like, and can optimize out proper algorithm parameters according to different hydroelectric generating set models in reality, and the identification precision is effectively improved through the improved algorithm.