Recent improvements in the actuators and the control methods allow the production of humanoid robots that may outperform human beings in balance keeping control under the effect of suddenly occurring disturbances. For this purpose, the current study proposes using the model predictive control (MPC) law in order to provide robustness for a humanoid robot against sudden disturbances. The main motivation for choosing MPC is that it naturally provides precautions against the predicted future disturbances and parameter changes. In order to check the performance of the humanoid robot considered here in a balance recovery reaction under the effect of a sudden external disturbance, its performance is compared to the experimentally recorded responses of a human subject. A single-axis tilt-platform is used to apply the sudden disturbance experimentally to the human subject. The results of the experimental runs show that there are obviously noticeable differences between the initial and final postures of the human subject. Therefore, the MPC method is applied to the humanoid robot in two different versions. In the first version, the robot tries to come back to the initial equilibrium posture. In the second version, the robot aims at the final equilibrium posture, which is observed in the experimental data of the human subject. The automatic prediction of the final equilibrium posture by the robot itself without imitating a human being will possibly be a typical subject of interest in the forthcoming studies. In this study, the robot detects and predicts a suddenly applied external disturbance automatically, and it generates the necessary control inputs to recover its balance. The simulation results indicate that the humanoid robot with MPC can give a much faster and smoother balance recovery response to the same external disturbance as compared to the response of the actual human subject. Hence, it is concluded that the proposed MPC law may be applied successfully to the humanoid robots that will relieve the human beings from several risky-environmental tasks in the future.