Abstract— In the near future, the humanoid robot has been expected to associate and work with a human. There is a chance that it has been hit from an external force and the robot cannot keep its balance. Thus, the robot might falls to human causing casualty or if it falls down to the ground the damage could cause ultimately to itself. For this reason, the humanoid must have balance recovering processes for protecting itself from the external force to prevent such damage. Therefore, this research proposes an optimal path design for a stand-balancing humanoid robot. The experiment simulates this situation using a force 1.11N hits to the humanoid (Bioloid Premium Type A) robot. This commercial humanoid robot has 18 Degree of Freedoms (DOFs). With this complexity of DOFs, the mathematical model and joints control strategies are investigated to restore robot balancing. Six strategies are chosen to implement in this work; 1) ankle strategy, 2) knee strategy, 3) ankle and knee strategy, 4) ankle and hip strategy, 5) ankle knee and hip strategy, and 6) whole body (ankle, knee, hip, arms) strategy using Multi-objective Whale optimization algorithm (MOWOA) together with non-dominated solution and decision making by weighted product method. Three objective functions are employed; 1) a minimal orbital energy, 2) a minimal error of phase portrait, and 3) a minimal jerk. The results have shown that the ankle strategy gives the best result based on decision making by the weighted product method.