There has been a tremendous growth over the past decade in the use of wireless communication. As the cost of wireless access drops, wireless communications could replace wired in many settings. Today, widely travelling laptop users access the Internet at a variety of places including their homes, and even at public places such as airports. Mobile Wireless Ad hoc Networks (MANET) is one such type of wireless network that have many useful applications including Wireless Sensors Networks which is now used in many civilian and environmental application areas. In mobile ad-hoc networks, nodes act as both routers and terminals. For the lack of routing infrastructure, they have to cooperate to communicate. Misbehaviour means deviation from regular routing and forwarding. It occurs by either selfish or malicious nodes. In both types misbehaviour's impact on MANET's proves to be detrimental, decreasing the performance and the fairness of the network, and in the extreme case resulting in a non-functional network. In this thesis we have addressed the requirements that nodes misbehaviour detection solution in MANET's should achieve. Existing solutions related to nodes misbehaviour detection in MANET were shown to fail to meet all of our requirements. The main direction of our work has been to look for an effective approach that can satisfy our requirements. The result is a new novel low cost framework entitled Sessions-basedM isbehaviour Detection Framework (SMDF). It consists of three components, the detection component, the decision component and the isolation component. We analysed and evaluated the proposed schemes by simulation techniques. By comparing our results to those of other mechanisms available in the literature, we showed that our solution has low cost in terms of communication overhead and has the lowest False Positive as well as the highest value of True Positive Detection Rates. It also showed that our solution has lower energy consumption rate and is scalable. Finally we present a series of proposals for future research work that have been raised by this work, such as tackling detection complications in hybrid ad hoc network environments.