In prior research,the orientation of head of the snake-like robot is changed according to the sinusoidal wave. To solve this problem,we propose Central Pattern Generator( CPG)-based control scheme with head-controller to stabilize the head of the underwater snake-like robot. The advantage of the CPG-based control scheme with head-controller is that the head of the underwater snake-like robot is direct to the target orientation during swimming. The relation between CPG parameters and orientation stability of head is discussed.The adaptation of the proposed method to environment changes is tested. The influences of CPG parameters and hydrodynamic forces on the orientation offset of head are investigated. The target orientation( the input of headcontroller) with an experimental optimization is calculated through a convenient method. To prove the feasibility of the proposed methodology,the different swimming modes have been implemented in our simulation platform.The results show that the oscillation of head’s orientation is inhibited effectively,and the proposed method has strong adaptation to environment and CPG parameters changes. In prior research, the orientation of the head of the snake-like robot is changed according to the sinusoidal wave. To solve this problem, we propose Central Pattern Generator (CPG) -based control scheme with head-controller to stabilize the head of the underwater snake-like robot. The advantage of the CPG-based control scheme with head-controller is that the head of the underwater snake-like robot is direct to the target orientation during swimming. The relation between CPG parameters and orientation stability of head is discussed The adaptation of the proposed method to environment changes is tested. The influences of CPG parameters and hydrodynamic forces on the orientation offset of head are investigated. The target orientation (the input of headcontroller) with an experimental optimization is calculated through a convenient method. To prove the feasibility of the proposed methodology, the different swimming modes have been implemented in our simulation platform. The results show that the o scillation of head’s orientation is cured effectively, and the proposed method has strong adaptation to environment and CPG parameters changes.