The trend of globalization motivates people to travel more often to different cities. In order to provide better suggestions for travelers, it is important to understand their preferences for venue types. In this article, we investigate travelers’ preferences based on the check-in data collected from a popular Location-Based Social Application called Swarm. We conduct a thorough analysis of the check-in data to discover the variation in travelers’ preferences between cities with different characteristics, and to build a model for predicting the venue types of travelers’ interests in each city. Our experimental results demonstrate that the F1-score increases by 0.19 by taking into account the characteristics of the destination city. Moreover, our approach outperforms collaborative filtering, a widely used approach to the design of recommendation systems.