Current gaits modification strategy for snake robots to adapt to environment is complex and needs long time to adjust control parameters. By introducing neural stepping stimulation mechanism, a simple and fast biomimetic control strategy is proposed based on the multi-phase central pattern generator (CPG). The multi-phase CPG which produces various snake robots gaits is built. Based on biological study, the neural stepping stimulation mechanism is proposed. Through the kinematic analysis of three major snake robots gaits, the relationship between control parameters and motion performance is obtained. Combined with snake robots locomotion characteristics, the complete snake robots adaptable biomimetic control strategy is built with the neural stepping stimulation mechanism. There is a simulation comparing the performance of this strategy with traditional control method for snake robots and a experiment verifying the control strategy.