First we use the probability statistics method to extract the characters from the artificial DNA sequences whose categories are known. Thu s we can get the character vectors of the DNA sequences and input them as sample s into BP neuron NN for learning.