When we encounter a devastating earthquake disaster, we have upgraded the earthquake resistant design codes in the long history of earthquake structural engineering. However the repetition of this action does never resolve the essential problem. This is because building structures and input ground motions have various complex uncertainties and unexpected phenomena often occur. The 2011 off the Pacific coast of Tohoku earthquake also provided some unexpected phenomena. This review paper discusses how to narrow unexpected issues in future earthquakes by referring to several concepts. Critical excitation methods, info-gap theories for uncertainty representation and interval analysis methods are the principal concepts.