Considering the nodes with different anti-attack abilities in scale-free networks，we investigated the probabilistic behaviors of malware propagation in scale-free complex networks. Using the cellular automata，we proposed a model of malware propagation in complex networks with the nodes having different anti-attack abilities. In particular，a vulnerability function related to node’s degree is firstly introduced into the model to describe the different anti-attack abilities of nodes. Then，the epidemic threshold and time evolution of malware propagation are investigated through analysis and simulation for the various vulnerability functions. The results show that different anti-attack abilities of nodes can produce significant effects on the behaviors of propagation. For example，different anti-attack abilities of nodes can change the value of epidemic propagation，and slow down the spreading speed of malware. Finally，it is pointed out that the vulnerability function is very important for adopting appropriate immunization strategies to control the malware propagation.