This paper deals with the application of segment procedure neural network to predict a chosen stock price. A novel procedure neural network is proposed to solve those problems which are related to some different segments of procedure. It is indicated that this model is a generalized form of the known procedure neural network, and it has all properties of the procedure neural network. This paper also presents learning algorithms for the segment procedure neural network. Stock price prediction is sophisticated because of its randomicity and non-disciplinarian, but it behaves as a typical segment procedure and has some inherent rules in each segment. This paper proposes a segment procedure neural network to model this issue, and presents some simulating experiment results.