Abstract: Introducing the knowledge relevant to fuzzy mathematics and microeconomics, an ABC-supported QoS unicast routing scheme is proposed. In the scheme the intervals are used to describe users' flexible QoS requirements with the edge-suited membership function introduced to adapt to the inaccurate information on the status of link. With the mechanisms introduced in for bandwidth pricing and edge/path evaluation and based on the ant algorithm, the QoS unicast path is sought to enable the utility of both users and network provider to achieve or approach the Pareto optimum in Nash equilibrium. The routing scheme was implemented through NS2 simulation, and its performance was evaluated on several actual and virtual network topologies. The results showed that the routing scheme is feasible and efficient.