Abstract: This study begins by constructing the Investor Climate Sentiment (ICS) Index for China using the Naive Bayes algorithm. It then applies the Nonlinear Thermal Optimal Path (TOP) method to examine the dynamic relationship between ICS and the volatility of China’s crude oil futures market. Finally, leveraging the dynamic results from both the Extended Heterogeneous Autoregressive (HAR) model and the TOP model, the study presents a dynamic and accurate prediction of crude oil volatility. The empirical findings indicate a significant dynamic leading relationship between ICS and crude oil futures prices, with this relationship becoming progressively stronger. In-sample regression results from the HAR model demonstrate that ICS has a long-term effect of reducing market volatility, and the HAR model incorporating ICS shows a strong fit. Furthermore, in out-of-sample volatility forecasts, the predictive power of the model with the ICS index is significantly enhanced.