Abstract: On the basis of Bai-Perron multiple structural mutation test from the nonlinear perspective，this thesis used GARCH models to re-research on the volatility and“leverage effect”of the Chinese stock market.The research finds：（1）in the sample period，structural mutation took place twice and once on the Shanghai and Shenzhen stock markets respectively：Feb. 1，2005 and Oct. 16，2007 in Shanghai and Oct. 31，2007 in Shenzhen；（2） the volatility and“leverage effect”before and after the structural mutation varied greatly either in the Shanghai Stock Exchange or Shenzhen，but the general characteristics of the Shanghai and Shenzhen markets had some similarities，especially in the last sample period after structural mutation in October 2007：both markets had no risk premiums but there was a significant asymmetrical“leverage effect” and bad news would give investors a greater impact than good news；（3）further research of the same sample period found，there was a significant coupling effect between Shanghai and Shenzhen’s stock price indexes but not between their yield sequences.

Key words: Shanghai and Shenzhen stock markets, volatility, structural mutation, leverage effect, nonlinear, GARCH models