Exquisite controller of nervous system, so-called, a central pattern generator (CPG), exists in rhythm phenomena of organism. Mathematical network models were proposed to account for the phenomena, however, the physiological details has not been clear. The basic form of this model is a combination of reciprocal inhibition of nervous systems that control extensor and flexor, respectively, and, as such, must be observed mutual bursting discharge. Rybak et al. proposed a Hodgkin-Huxley type mathematical CPG model composed of two levels.We constructed, in this study, reproduced quadruped walking patterns (gaits) by means of coupling four hardware CPG models, based on Rybak's mathematical model. So far, such the networked CPG model was simultaneously controlled by plural parameters to generate abundant gaits, or network structure itself was changed every gait. In this study, we proposed a hard-wired CPG network that has been possible to generate two gaits, by changing just one parameter.