With the rapid development of internet, the technology of email is widely used in people's daily life. However, the occurrence of more and more spam emails is annoying to user, which causes the great waste of user's time, money as well as network bandwidth. And what's worst, it can be harmful to users. For example, pornographic content may be contained in spam email. Therefore, it is very important to resolve the problem of spam emails.Based on the systematic summarization of the most recent work on anti-spam, this dissertation explores the techniques of anti-spam email. Through the author's effort, some innovations and achievements are made by the author, which will be illustrated in detail as follows.The breakthrough of this dissertation could be embodied in the following several respects:1. This dissertation summarizes new characteristics of present spam. In order to avoid filtering, more and more spam emails use new tricks so that many spam filteres have not been able to work effectively. Based on studying a large number of papers and researching spam emails, the dissertation summarizes theory and characteristics of new tricks.2. This dissertation presents a content-based MNNB (Markov N-gram Naive Bayes) algorithm to anti-spam. The algorithm uses Markov chain to improve disadvantage of Naive Bayes algorithm, which assumes attributes are conditionally independent. Experiments show that MNNB algorithm has better performance than Naive Bayes algorithm and is more available for filtering spam email of different languages.3. This dissertation presents a novel anti-spam e-mail algorithm based-LVQ network in terms of spam e-mails which are mainly made up of several kinds commercial or political spam emails at present. As LVQ neural network may create complex boundaries through combining several