Abstract: Markov-switching model is a method applied to investigating the structural changes of time series. To explore quantitatively characteristics of the stock market of China, the Markov-switching model in the form of three states, heteroskedasticity and fourtlh-order autoregression is built to analyze the wave of the component index of Shenzhen Stock Exchange. And some characteristics of the stock market of China are summarized at last.

Key words: Markov-switching model, stock market of China, autoregression, heteroskedasticity