Handoff is a call handling mechanism that is invoked when a mobile node moves from one cell to another. Such movement may lead to degradation in performance for wireless networks as a result of packet losses. A promising technique proposed in this thesis is to apply multicasting techniques aided by mobility prediction in order to improve handoff performance. In this thesis, we present a method that uses a Grey model for mobility prediction and a fuzzy logic controller that has been fine-tuned using evolutionary algorithms in order to improve prediction accuracy. We also compare the self-tuning algorithm with two evolutionary algorithms in terms of accuracy and their convergence times. Our proposed method takes into account signal strengths from the base stations and predicts the signal strength of the next candidate base station in order to provide improved handover performance. The primary decision for mobility prediction is the accurate prediction of signal strengths obtained from the base stations and remove any unwanted errors in the prediction using suitable optimisation techniques. Furthermore, the model includes the procedures of fine-tuning the predicted data using fuzzy parameters. We also propose suitable multicasting algorithms to minimise the reservation of overall network resource requirements during handoff with the mobility prediction information. To be able to efficiently solve the problem, the situation is modelled using a multicast tree that is defined to maintain connectivity with the mobile node, whilst ensuring bandwidth guarantees and a minimum hop-count. In this approach, we have tried to solve the problem by balancing two objectives through putting a weight on each of two costs. We provide a detailed description of an algorithm to implement join and prune mechanisms, which will help to build an optimal multicast tree with QoS requirements during handoff as well as incorporating dynamic changes in the positions of mobile nodes. An analysis of how mobility prediction helps in the selection of potential Access Routers (AR) with QoS requirements - which affects the multicast group size and bandwidth cost of the multicast tree -- is presented. The proposed technique tries to minimise the number of multicast tree join and prune operations. Our results show that the expected size of the multicast group increases linearly with an increase in the number of selected destination AR's for multicast during handoff. We observe that the expected number of joins and prunes from the multicast tree increases with group size. A special simulation model was developed to demonstrate both homogeneous and heterogeneous handoff which is an emerging requirement for fourth generation mobile networks. The model incorporates our mobility prediction model for heterogeneous handoff between the Wireless LAN and a cellular network. The results presented in this thesis for mobility prediction, multicasting techniques and heterogeneous handoff include proposed algorithms and models which aid in the understanding, analysing and reducing of overheads during handoff.