With the complex shifts in the international situation, global supply chain networks (SCNs) are now facing disruption risks. The primary objective of this study is to explore the evolutionary mechanisms of multi-level supply chain networks (MLSCNs) and simulate the risk propagation throughout the network. Firstly, this study formulates a MLSCN influenced by the combined preference. It contemplates a cascading risk propagation model that accommodates for node perturbations. Secondly, this study proceeds to examine the effects of parameter variations and capacity dynamics on risk propagation. Finally, the effect of failure probability on the robustness of MLSCNs is considered in the case of node overloading. The conclusions indicate that the developed framework fosters a more robust topology within MLSCNs. Within the context of the cascading risk propagation model, it is observed that parameter variations have a significant impact on the cascading propagation of risk. Moreover, the regulation parameter for load reallocation effectively slows down the rate at which risk cascades to network failure. The study also reveals that dynamic changes in node capacities can intensify network vulnerability, and make it more prone to collapse. Failure probability emerges as a potential countermeasure to prevent network collapse. The outcomes of this research can provide a valuable reference for both researchers and business managers to better understand SCNs and the cascading propagation of risks within them.