This paper deals with the whole body motion control problem of humanoid robots by using automatic and bilateral control methods. Advanced motion controllers are proposed so that human daily life activities, such as walking on an uneven terrain and contacting an unknown object, can be performed. It is shown that the uncertainty of environmental impedance variation is one of the main challenging issues in the design of the automatic control systems, i.e. a humanoid robot may easily lose its stability when it contacts an unstructured environment. A novel bilateral control method is proposed for humanoid robots so as to overcome the environmental uncertainty problem. It is shown that more dexterous and versatile tasks can be easily performed by transforming the skills of humans to humanoid robots in the proposed bilateral control system. A disturbance observer is used not only to achieve the robustness of the motion control system but also to perform sensorless contact motion control tasks. The performance of sensorless contact force estimation is significantly improved by deriving the exact dynamic model of humanoid robots. A floating point base exact dynamic model is derived by using the extended Newton Euler algorithm. A new robot simulator is designed by using the proposed dynamic model and the Virtual Reality toolbox of MATLAB. Simulation results are given to show the validity of the proposals.