Abstract: Along with the development of civil aviation in our country,the airport noises are getting more and more serious．Aimed at the airport-noise time series prediction problem,this paper presented the prediction model based on SSA,in which the airport-noise time series are obtained by singular value decomposition,and the principal component and the Empirical orthogonal function are obtained,the characteristics of trend and vibration are analyzed and then the appropriate feature vectors are selected for sequence reconstruction,and prediction model is constructed by linear repeating formula．Based on the state-transition matrix and contribution ratio,the forecast values are revised．The experiment on the measured data of an airport shows that the accuracy of this model is better than other original prediction models.

Key words: SSA,Airport-noise time series,Prediction model,State-transition matrix,Contribution ratio