The modern statistical methodology and practical statistical sciences have been systematically generated since the establishment of the grammar of science for recognition to give laws to Nature in 1892 by K. Pearson. The author proposes to construct a new grammar of science for design and corresponding new mathematical or statistical methodology to improve the laws for society such as the Taguchi method. In this new grammar, not passive facts based on the researchers’ sense impression but active facts designed by numerical simulation will take an essential role. The author also introduces recent research activities in Japanese Society for Quality Control to establish new statisical methodology for technology developments based on numerical simulation.