With the popularity of social networks

social worms have constituted one of the major security threats.This type of worm spreads fast in Internet networks based on topology information and social engineering.There are two critical problems of previous works

which result in underestimating the scale of infected network and lacking a comprehensive and accurate abstraction for human behaviors.These are structural imperfection of network topology and one-sidedness of the propagation model.To address the above problems

this paper proposed a novel simulation model of social worms.We more accurately and comprehensively abstracted the relationship between social logical layer and actual physical layer in a hierarchical network

and characterized the social worm propagation through the temporal characteristic of human mobility.Experimental results show that the impacts of user behaviors

topology parameters and different recovery processes on the propagation.Meanwhile

we give the qualitative analysis of the spread ability.