When visiting a new city, tourists often need help to identify personally interesting places or locations from a potentially overwhelming set of choices. Recently, the increasing availability of GPS-enabled devices and the rapid advances in geotagged social media have led to the accumulation of a large amount of location histories, which may reflect people-s travel experiences in the environment. Research has shown that experiences from past users (especially similar ones) in similar contexts can help the current users to solve their problems efficiently, e.g., choosing where to visit next. Motivated by the above aspects, this thesis explores a methodology of deriving recommendations from location histories in Location Based Services (LBS). More specifically, we investigate collaborative filtering methods for deriving personalized and context-aware location recommendations from human location histories (e.g., GPS trajectories and trajectories constructed from Flickr photos). These methods can be implemented in LBS to provide tourists with location recommendations matching their interests and context when visiting a new environment (e.g., city or museum). The insights gained in this research can be also transferred to many other applications, such as friend recommendations in location-based social networks, artwork recommendations in museums, recommendations in the shopping domain, human behavior understanding, and activity recognition.