A method is proposed to find key components of traffic networks with homogenous and heterogeneous topologies, in which heavier traffic flow is transported. One component, called the skeleton, is the minimum spanning tree (MST) based on the zero flow cost (ZCMST). The other component is the infinite incipient percolation cluster (IIC) which represents the spine of the traffic network. Then, a new method to analysis the property of the bottleneck in a large scale traffic network is given from a macroscopic and statistical viewpoint. Moreover, three effective strategies are proposed to alleviate traffic congestion. The significance of the findings is that one can significantly improve the global transport by enhancing the capacity in the ZCMST with a few links, while for improving the local traffic property, improving a tiny fraction of the traffic network in the IIC is effective. The result can be used to help traffic managers prevent and alleviate traffic congestion in time, guard against the formation of congestion bottleneck, and make appropriate policies for traffic demand management. Meanwhile, the method has very important theoretical significance and practical worthiness in optimizing traffic organization, traffic control, and disposal of emergency. One method, is the minimum spanning tree (MST) based on the zero flow cost (ZCMST). The other component is the infinite incipient percolation cluster (IIC) which represents the spine of the traffic network. Then, a new method to analyze the property of the bottleneck in a large scale traffic network is given from a macroscopic and statistical viewpoint. Moreover, Three effective strategies are proposed to alleviate traffic congestion. The significance of the findings is that one can significantly improve the global transport by enhancing the capacity in the ZCMST with a few links, while for improving the local traffic property, improving a tiny fraction of the traffic network in the IIC is effective. The result can be used to help traffic managers prevent and alleviate traffic congestion in time, guard again st the formation of congestion bottleneck, and make appropriate policies for traffic demand management. And the appropriate method for traffic demand management. Meanwhile, the method has very important theoretical significance and practical worthiness in optimizing traffic organization, traffic control, and disposal of emergency.