Next generation telecommunication networks are increasingly turning to value-added services in order to attract more users and increase revenues. This stimulates development of new telecommunication value-added services which tend to rely on user context for enabling enhanced service capabilities. There are multiple aspects of user context and knowledge of one or more such aspects enables service’s user profiling. One possible aspect of user context is the user location information, and services that use such information are called location services. However, knowing just the location of a user no longer meets the needs of new services and there is a need for additional insight into the user location information. One possible extension to location information is insight into user movement. With insight into user movement it may be possible to predict it as well, which opens up a whole array of new services and functionalities. To enable handling of such location information to value-added services, an adaptive location management model is proposed. The proposed model includes the tasks of location interpretation, prediction of user or group movement and creation of groups based on similarities in the user movement. Regarding the purpose of the model we propose its inclusion in the next generation network’s service stratum, which holds the functions related to services. The inclusion of the proposed model is based on an entity responsible for ensuring the necessary adaptive location management functions. Placing such entity in the next generation network service stratum enables value-added services to use entity functions regardless of services being in or outside the network itself. We propose several location management functions that can adapt their functionalities according to service requirements. Services are able to customize provided location accuracy and interpretation, precision of the movement prediction and definitions of user movement similarity for the purpose of user grouping. Movement prediction is the most important aspect of location management provided by the proposed model. Therefore, the thesis proposes several prediction models whose execution is depending on the knowledge of user context. In this sense, movement prediction can be performed based on knowledge of user movement history and based on knowledge of the environment in which the user is moving. For the purpose of movement prediction artificial neural networks and probabilistic models are used. Along with movement prediction, the thesis proposes trajectory, destination and time related location prediction, which are derived from the initial movement prediction models. In order to be able to predict user movement based on movement history we propose analytical procedures used for movement regularity detection. In addition to prediction models, the thesis proposes models for location scaling and user grouping. Movement prediction models, location scaling and user grouping procedures are evaluated on value-added service example. A paperless ticketing service is used for this purpose and it is shown how adaptive location management can benefit such services. The evaluation of prediction models is performed on real mobile network users’ data.