The feedforward neural network is the most fundamental type of neural networks from historical viewpoint and its wide applicability. This chapter discusses several aspects of this type of neural network in detail. Section 2.1 describes its fundamental structure and algorithm, Sect. 2.2 various types of layers, Sect. 2.3 some techniques for regularization, Sect. 2.4 the acceleration techniques for training, Sect. 2.5 the methods for weight initialization, and finally Sect. 2.6 the model averaging technique and the Dropout.