The digital characters recognition method based on the combined feature which is used as input layer feature vector in Bp neural network is proposed. First, segment feature, which is introduced in this paper, and the ranks of the statistical characteristics of the numeric characters are combined as a new feature. Then, Bp neural network classifier is designed according to the combined new feature. Finally, when the training sample set is trained according to the training method of the BP neural network, the recognition results of the test sample set can be acquired. According to the result, the recognition rate can reach at above 94%.