In this article, we build a nonlinear dynamics model of credit risk contagion in CRT market based on the connections of the complex relationship network between participants in CRT market only include Newman-Watts length scale connection and long distance connection, and study its dynamic behaviors. We find that, firstly, the increase of the effective rate of credit risk contagion can make the status curve of credit risk contagion happen some significant changes, further engender the Hopf bifurcation and chaos phenomenon in the process of credit risk contagion through experimental simulation. Secondly, the increase of the nonlinear resistance coefficient make Hopf bifurcation and chaos phenomena happen in advance. Thirdly, there are series of periodic windows in the chaos interval inside, in which present three intertwined state about Hopf bifurcation, pour bifurcation and chaos.