To achieve QoS, independently of the routing protocol, each mobile node participating in the network must implement traffic conditioning, traffic marking and buffer management (Random Early Drop with in- out dropping) or queue scheduling (Priority Queuing) schemes. In MANETs, since the mobile nodes can have simultaneous multiple roles (ingress, interior and destination), it was found that traffic conditioning and marking must be implemented in all mobile nodes acting as source (ingress) nodes. Buffer management and queue scheduling schemes must be performed by all mobile nodes. By utilizing the Network Simulator (NS2) tool, this thesis focused on the empirical performance evaluation of the QoS schemes for different types of traffic (FTP/TCP, CBR/UDP and VBRI/UDP, geographical areas of different sizes and various mobility levels. Key metrics, such as throughput, end-to-end delay and packet loss rates, were used to measure the relative improvements of QoS- enabled traffic sessions. The results indicate that in the presence of congestion, service differentiation can be achieved under different scenarios and for different types of traffic, whenever a physical connection between two nodes is realizable.