In an anonymous secret sharing scheme, the secret can be reconstructed without knowledge of which participant hold which share. That is, in such scheme the secret can be recovered from the shares without the identities of shareholders. However, the previous anonymous schemes are not very efficient and the existing bounds make optimization work difficult.In this paper, we present a novel anonymous secret sharing scheme by using BP Artificial Neural Network. The scheme can ensure that the shareholders' identities are completely anonymous when the secret is recovered. Unlike previous schemes, our scheme is an ideal (t, n) threshold scheme which parameter t doesn't have lower bounds and need less storage space. Furthermore, the shares can be chosen by shareholders and the communication between dealer and participants doesn't need secure channel. In addition, the participants can verify the correctness of their shares and the secret S.