Seismic risk for cultural heritage buildings has been underestimated for years in Switzerland. A strong earthquake can occur at any time in this region, as it has been shown during the last centuries. As reminder the 1356 seismic event that destroyed the city of Basel can be quoted; this event, whose intensity is assessed as having been I=IX on the European Macroseismic Scale (EMS-98), is recorded as the most violent earthquake that struck central and northern Europe. Since 2003 or even 1989, the seismic safety of common buildings is well defined through modern building codes (Swisscodes (SIA)). However, this is not the case of cultural heritage buildings whose seismic vulnerability has been only partly addressed. This PhD thesis was initiated in order to fill in this void of knowledge by creating a methodology for assessing the seismic vulnerability of historical edifices. Surveys carried out on cultural heritage buildings after seismic event showed that such kind of structures are particularly vulnerable under seismic actions. This vulnerability is essentially due to their particular structure that is characterized by slender components (pillars), wide open spaces, quite big masses located at high levels (vaults and their filling, lantern towers, etc.) on one hand; on the other hand, they generally have few components that can resist the lateral actions perpendicular to the nave and their masonry fabric is non-ductile. Moreover the non-existence of a stiff horizontal component (or a deformable one in case of wooden ceilings) and heavy masses concentrated in walls rather at floors make it impossible to apply models that have been developed for common buildings. The methodology that is presented in this report is composed of four steps. It gives the possibility to first sort out the sacred edifices that are seismically vulnerable from the ones that are not vulnerable. The second step allows engineers to determine the seismic vulnerability of a given edifice by the use of simplified models that give valuable results. In case more accurate results are required, more sophisticated models can be applied, as the Finite Element Method. In the last step, the seismic vulnerability of the edifice is put in parallel with the expected peak ground acceleration of the region. This last phase permits to set if the given edifice can be damaged by earthquakes, which damage is expected and which retrofitting is required.