In this article bio-inspired techniques are used for generating the gait of a biped robot. The concept of CPG, central pattern generator, which is a neural network capable of producing rhythm output, was used. It was modeled as coupled oscillators. With the purpose of verifying the operation of the oscillators, simulations were made. After that they were implemented for generating the gait for the robot. The output of the oscillators were used as the reference trajectory for the feet and the joint angles were obtained by inverse kinematics. We obtained good result in simulations for the humanoid robot, which showed a degree of tolerance to external perturbations.