Wireless Sensor Networks (WSN) have an important role to play in applications involving surveillance, security and autonomous systems. Furthermore, recent technological advances have allowed wireless sensor networks to be applied to a plethora of areas such as environment monitoring, traffic control, health, agriculture, medical, home applications, as well as fire fighting, and object tracking. One of the main, generic WSN requirements is the collection of large amounts of data which can be afterwards used in classification and decision making processes. Within such a general WSN framework, this dissertation studies sensor node positioning strategies. Thus given a fixed number of sensors operating in a completely unknown environment, work is focussed on the development of efficient sensor positioning techniques. Efficiency here relates to i) collection of data in order to characterize (for a given accuracy) the environment, with a minimum number of sensor moving steps i.e. as quickly as possible, and ii) the location and tracking of major features of the environment, for example the maxima of data distributions used to form in simulations the data profile of a given environment. Furthermore, the above WSN movement/positioning methodologies are applied to both data static and data dynamic environments. Note that these methodologies contain two key processes: i) data interpolation; and data prediction as applied to trajectories of moving environment features. Thus WSN data is interpolated using a form of Kriging interpolation whereas prediction is performed using a polynomial based approach. Experimentation has been performed using computer simulation of proposed methods and experimental results are presented in the thesis which allows proposed schemes to be compared in terms of different criteria. Results associated to systems employing ground truth data, as a substitute for ideal interpolation and prediction processes, are also presented and are taken as providing bounds of system performance.