Robot are expect to be new tools for the operations and observations in the extreme environments where human has difficulties for accessing directly, deep ocean, space, nuclear plants and so on. One of the important matters to realize mobile robots for extreme environments are to establish systems in the movement and their structures which are strong enough to disturbance. A solution for realization an adaptive control system is to learn and imitate biological systems. For example, in the spinal cord of animals, neural oscillator systems called Central Pattern Generator (CPG) are proven to exist and investigated that CPG control rhythmical signals such as swimming pattern, walking locomotion, heart beats, etc. In this paper, an adaptive control system based on the entrainment feature of neural oscillators was developed and applied to motion control of a snake-like robot.