Automated negotiation is a key form of interaction in agent-based systems. One-to-many negotiation, one of the typical negotiation modes, which can be treated as multiple, concurrent one-to-one bilateral negotiations, is discussed as well as its coordination strategy in this paper. A flexible one-to-many negotiation system is proposed. Compared with the existing general one-to-many negotiation systems, buyer agent should not wait until having received offers from all its trading partners before generating counteroffers. Also each seller agent can join in or leave the negotiation dynamically. Negotiation process is open, dynamic and continuous. The coordination strategy based on relative utility is discussed that can improve the satisfaction degree of whole negotiation when more than two same maximum utilities are reached among several concurrent one-to-one negotiations. The experiment shows that the time cost and the agreement are optimized base on the system proposed in this paper.