AbstractWith the rapid development of autonomous driving technology and the sharing economy, future transport systems may be a mixed traffic environment of conventional travel modes and autonomous vehicles. This study designed a travel survey with cognitive experience for autonomous vehicles to obtain the stated preference data of different commuters for multimodal travel choices. Then, discrete choice models were established, and sensitivity analysis was conducted to deeply analyze commuters’ choice preferences and heterogeneity in the multimodal travel environment. It is concluded that mixed logit models incorporating cognitive and attitudinal variables are more suitable for discussing travel choice behavior for autonomous vehicles. In addition to factors related to travel modes and personal information, cognitive experiences and attitudes are the two main determinants of multimodal travel choices regarding autonomous vehicles. Increasing travelers’ cognitive levels and inculcating attitudes toward autonomous vehicles through advertisements and experiences can encourage more commuters to choose private or shared autonomous vehicles. Private car commuters mainly switch to private autonomous vehicles, and public transport commuters mainly switch to shared autonomous vehicles with increasing cognitive change and inculcating attitudes toward using autonomous vehicles. The research results can provide a reference for travel behavior analysis and traffic policy formulation for different commuters in the era of autonomous vehicles.