When massive information brings people more channels to attain messages, there generates different types of new personalized recommendation systems. As the important sector for social development, tourism industry suffers the problem of information over-loading. In the construction of personalized recommender system mainly involves two problems: information acquisition and personalized recommender. This recommender system is able to form client’s databank after user’s login and assessment for various travel destination and products to support more accurate user’s information mining. The authors adopt improved hybrid recommender algorithm through combining collaborative filtering algorithm with content-based recommendation algorithm. The relationship of user and travel destination can be classified as the interested and disinterested. So it can predict the degree of new users’ enthusiasm towards different types of travel destination to realize personalized travel recommendation.