Abstract: This paper mainly introduces the comparative study of methods dealing with the spam in receiving end, which includes preprocessing, feature selecting and classifying. Documents frequency, information gain and odds ratio are all the methods of feature selection.This paper also introduces a new method of feature selection,which is knowledge gain. Its excellent behavior is veriifed in the experiment. Common classiifers include KNN, Bayes, SVM, etc. The liner classiifers also have the advantages in spam which is presented in the experiment too.