The invention discloses a wind power ultra-short term probability prediction method and system based on empirical dynamic modeling, and the method comprises the steps of carrying out the standard normalization processing on a time sequence of a to-be-predicted quantity, and carrying out the nonlinear aggregation degree calculation of data after the standard normalization processing, so as to inspect the nonlinear degree of a given dynamic system; calculating an optimal embedding dimension E and delay time tau by adopting a particle swarm optimization algorithm; further, performing phase spacereconstruction on the time sequence of the to-be-predicted quantity; and constructing an empirical dynamic model, and predicting the given dynamic system in the reconstruction phase space by adoptinga simplex projection method to obtain a prediction result of the to-be-predicted quantity. The prediction result shows that the wind power ultra-short term probability prediction method based on empirical dynamic modeling can achieve objective description of the wind power generation dynamic process completely according to data, and the effectiveness of probability prediction is remarkably improved.