In this paper, standard SIR model of rumor In this paper, standard SIR model of rumor spreading is used in the weighting complex networks. The tie strength is expressed by the function of edge weight\textit f(ω) in order to simulate the function between relationship and reliant interpersonal. In the practical scene, a node may only be contacted by some of its neighbors, so that the nonlinear rumor spread exponent α of Anurag Singh (2013) is introduced, and the rumor spreading in complex networks is discussed. The effect of \textit f(ω) and α on the rumor threshold is discussed in the paper. It is found that the mean edge weight<ω_kk'together with its function\textit f(ω) and μ are both effect the steady state value of r(t), that is \textitr(∞), which means the rate of people who heard the rumor in the crowd. And the explosion time is also effected by them. In the research, a positive tumor threshold λ_c had not been found when the \textit f(ω) and μ are introduced.