This paper proposes a gait planning approach to reduce the required friction for a biped robot walking on various surfaces. To this end, a humanoid robot with 18 DOF is considered to develop a dynamics model for studying various 3D manoeuvres. Then, feasible trajectories are developed to alleviate the fluctuations on the upper body to resemble human-like walking. In order to generate feasible walking patterns, not only horizontal interaction moments for the computation of ZMP, but also horizontal forces and vertical moment constraints between the feet and the ground surface are taken into account. Since the pelvis trajectory does drastically affect the walking pattern, the focus will be on generating a smooth motion for the pelvis. This smooth motion is generated based on a desired motion for the robot’s Centre of Mass (COM), which is mapped to the joint space using inverse kinematics. In fact, the proposed approach involves computing a moving ZMP based on a predefined desired COM trajectory to reduce the required friction for stable walking. The suggested gait planning approach (Low Friction Demanding Moving-ZMP, LFDM) is compared to various existing approaches considering slippage conditions. The obtained results reveal the effectiveness of the proposed method for various walking speeds which will be discussed.