Although many travel recommendation systems are worked with mobile social network, the articles and delivered messages between users are still not to be utilized well for estimating user preferences. If the articles and the delivered messages are closed related with the touring experiences, the information should be merged to the recommendation systems. Therefore, the information can be used for generating recommendation POIs through the intelligent modules. Besides, if the groups of social network and the exchanged information are collected large enough, the generating POIs should be more closed to user preferences. But the ideal situations are still not provided on the developed traveling recommendation system of current Internet. Therefore, we proposed a cloud-based Intelligent Touring System (ITS) to provide personalized PoIs instantly. By combining developed cloud computing technology, we can achieve analyzing posted blogs in mobile social network and recognizing user pedestrian patterns through smart phone sensors. Afterward, in order to generate the suggested PoIs related with user preferences, we construct meta-group based on similar user preferences. Moreover, in order to achieve meta-group classification, we developed clustering algorithm based on CLOPE . Finally, we conducted experiments for examining the performance of the proposed meta-group classification, and the results show it could construct groups with similar preferences and provide suitable PoIs to users during touring.