Traditional airport noise prediction models are insufficient for their high modeling cost and poor practicability. In this paper, the time series phase space reconstruction theory is introduced, and a novel integrated airport noise prediction model based on fast extreme learning machine and differential evolution is proposed. In the proposed model, the airport noise time series is reconstructed based on the phase space reconstruction theory, and the fast extreme learning machine is used to model the reconstructed phase space vector. Meanwhile, an improved differential evolution algorithm is adopted to search for the optimal parameter combination of phase space reconstruction parameter and model parameter simultaneously. The whole modeling process of the integrated prediction model is very simple and?efficient without any manual?intervention. Experimental results demonstrate that the proposed model can track the variation tendency of airport noise well and can achieve much more accurate prediction results than its counterparts.