Procedures of seismic hazard assessment are based on ground motion attenuation relationships which can be derived from statistical analyses of recorded ground motions or, in conditions of limited strong-motion records, can be obtained from the available literature sources. Ideally, these attenuation models should consider regional earthquake source and propagation path effects and local site response peculiarities. At present, these is no doubt that the relationships are different for different seismic regions, and “region, and site-specific” models should be developed. However, the empirical databases usually consist on the recording which were obtained during the events of small and moderate magnitudes. Therefore, a question is arising-“is it possible to use these relationships in the case of large events”? The lack of strong motion data close to large earthquakes results in using of several seismological models. An important criterion for any simulation procedure is its validation against recorded ground motion. Every strong earthquake provides an unique opportunity both to verify the accepted attenuation models, and to update empirical relationships.