Abstract:This paper builds an investor sentiment index by using data of Shanghai stock exchange,filters out its autocorrelation,and obtains the innovation of investor sentiments.We divide investor sentiments into posi_x005ftive and passive investor sentiment,and analyze the characteristics of investors’behavioral traits in different sentiment states.We then research the asymmetrical influence of positive and negative emotion on stock price behaviors by using Dummy Variable model,GARCH-model and RV-AR-model. The results show that in Chinese stock market,the models which consider different sentiment states have better fitting effects; positive investor sentiment has a significant impact on stock returns,but the effect of passive investor sentiment on stock returns is not significant since the rational component plays a leading role in the market when sentiment is low.Besides,the volatility of investor sentiment has a great explanation ability to the volatility of stock returns.