With the development of mobile networks

SMS spam has become an important challenge which need to be solved urgently.Current spam filtering technique is hard to fulfil the task accurately and timely in the presence of large-scale SMS systems.The paper proposes a mechanism based on collaborative behaviours and contents analysis and further develops a prototype system.In order to obtain a good filter result in time

we take the strategy of behaviour analysis

assisting with content analysis and introduce the probability theory into unitary short messages in Chinese.The proposed system improves the current SMS spam filtering methods by overcoming their disadvantages.By real SMS spam data running on the prototype system

we have showed that it can improve the effect of spam filter efficiently.