Recently, the concept of robust design using Taguchi method has widely been introduced in the fields of design optimization. Above all, robust optimal design takes into account the effect of noise factors in optimal design. This robust optimal design has been studied by several branches. The representative methods of robust optimal design are the ones using Taguchi method directly and one using the conventional sensitivity method. For robust optimal design using the conventional sensitivity method, conventional optimal design problem is re-formulated as the robust optimal design problem, and the process solving the robust optimal design problem is suggested. Finally, a comparative study on robust optimal design using Taguchi and sensitivity analysis methods is performed with two examples.