Abstract: Aiming at the shortage of traditional weighted network models such as BBV model that exhibit low clustering coefficient and weak clusteringdegree relation, an evolving model for weighted network with tunable clustering coefficient is proposed. In this model, the newly added node chooses the existing nodes to establish new edges according to preferential attachment scheme or triangle connection scheme, which is driven by node strength and initial attractiveness. Simulation results show that node degree, node strength and edge weight of the produced networks all obey the powerlaw distributions as initial attractiveness and triangle connecting probability get different values, and its clustering coefficient is tunable with these two parameters, especially, the relation between the average of clustering coefficient and degree exhibits better powerlaw relation when the triangle connecting probability gets larger values. Finally, a cascading failure model for weighted networks is established, and the influences of weights, clustering coefficient and other parameters on the networks’ cascading invulnerability are analyzed.