The effectiveness of "green travel" guidance is influenced by policies and travelers' choice preferences. Different types of low-carbon guidance policies and the heterogeneity of travelers need to be considered. To quantify the influence of policies and the heterogeneity of travelers on travel behavior, policy was categorized into three perceptions: economic, convenience, and public information policy perceptions. Based on the green travel questionnaire data of 521 residents, K-means clustering was used to classify the travelers into hedonic group (60.7%) and practical group (39.3%). The results of the partial least square method of structural equation model (PLS-SEM) suggest that in travelers' green travel behaviors, the effect of policy on travelers' green travel intention would be overestimated or underestimated if the heterogeneity of travelers is not considered. The green travel intention of the hedonic group is more affected by public information policy perceptions than practical group, and less affected by economic policy perceptions than practical group. Public information policy measures are the most important policy measures to promote green travel.