Abstract: Subjected to a variety of outside factors,change-point appears easily due to the imperfect legal system and market mechanism as well as psychologically immature investors and other reasons in the stock market of China. This paper determineds the locations of changepoints by introducing the method which is based on the theory of Bayesian inference,constructing a likelihood function,while using prior information and sample information,as well as adopting Gibbs sampling. In addition,Which considers the problem for the number of change-points as a problem of selection for the model,is determined by utilizing the posterior odds. After the introduction of the method,determined the number and locations of change-points in the sequence of Shanghai Composite Index. It was found in the study that,from January 1993 to May 2008,Shanghai Composite Index had four change-points as follows:about December 1994,January 2000 (or July 2001),April 2006 and November 2007.

Key words: bayesian inference, change-points, Shanghai composite index, gibbs sampling