Presynaptic and postsynaptic neurotoxins have very important application in basic research and drug design. The successful prediction of neurotoxin is becoming an important task in recent years. In this study, based on the concept of Chou's pseudo-amino acid compositions, an algorithm of increment of diversity (ID) is proposed for predicting presynaptic and postsynaptic neurotoxins. The results of jackknife test show that the accuracies of prediction are 90.23% for presynaptic neurotoxins and 89.40% for postsynaptic neurotoxins. In addition, toxins and non-toxins are also predicted by using this algorithm.