Support vector machine (SVM) is a kind of novel machine learning methods based on statistical learning theory,which has been developed to slove pattern classification problems.This paper applied support vector machine to chaotic time series analysis.The experimental data was Mackey-Glass chaotic time series.First,the support vector regression method was used on the chaotic time series regression prolbem.Then,Local-Region Multi-steps Forecasting Model was used with supprot vector machine to predict the chaotic time sereis.Simulation results show that SVM could predict the trend of chaotic time series correctly.