Slide drilling refers to the technology of creating a predetermined non-vertical wellbore with a bent housing positive displacement motor (PDM). It is widely adopted in the area of directional drilling. In practical drilling operation, the top drive on the ground introduces an angular rotation to the top of the drill string, and the PDM at the bottom of the drill string rotates accordingly. When the bend is pointed to the desired direction, the adjustment of the PDM stops and the drill string slides without rotation to make a deviation. Up till now, the relationship between the top drive displacement and the direction of the bend, namely the tool-face angle (−180° ∼ 180°), is still unclear. In this research, an indoor slide drilling experiment is carried out, and the nonlinear relationship between the top drive input and the tool-face output is recorded. The hysteretic phenomenon observed is consistent with the in-field experience, and a single-input-single-output (SISO) system is established to describe this relationship. The Volterra/ Wiener neural network (VWNN) is introduced to identify this system, and provides a one-step prediction of the tool-face output. The predicted tool-face output is verified by the experiment data.