The coefficient of risk compensation shows the compensation required by investors for taking each unit of risk，which also reflects the tradeoff between risk and return． Using data in the US stock market，this paper firstly builds a TVA-GAＲCH-M model to analyze the time-varying characteristics of investors’risk compensation coefficient． Then the relationship between the aggregate investor sentiment and the time-varying risk compensation coefficient in US stock market are investigated by using the granger causality test and the linear regression model． The results show that the coefficient of investors’risk compensation changes over time，and it is negatively related to aggregate investor sentiment． What’s more，the aggregate investor sentiment has a negative effect on the time-varying risk compensation coefficient． Particularly，investors with optimism sentiment will decrease their time-varying risk compensation coefficient while investors with pessimism sentiment will make their time-varying risk compensation coefficient increase．