The dissertation proposes the improvement of earthquake risk assessment methodologies for buildings.The research on which this work is based was initiated due to the knowledge that the earthquake risk in areas with moderate and low seismicity is not fully recognized and that the existing assessments of building vulnerability generate results of questionable reliability because they use global assessments of earthquake risk, exposure models and building typologies. Possibilities of improving the assessments of all three components of earthquake risk were explored. For the earthquake hazard, the advantages that the use of the probabilistic approach, uniform hazard spectra and microzoning of the observed area have on the final results of the earthquake hazard assessment have been investigated and proven. New attenuation equations were proposed for areas with moderate and low seismicity and also estimated were the values of horizontal and vertical peak accelerations for areas whose underlying soil is composed of deep layers of local soil and deep geological sediments. Seismic microzonation maps of the observed area were created. More reliable exposure modeling has been explored in terms of collecting data on buildings (and other risk elements) in site and considering only those building characteristics which are relevant to seismic vulnerability assessments. An exposure model of a representative city from the observed area was created, on the basis of which predominant typologies of buildings were proposed, for which it was determined that they could be assigned to buildings in all areas with similar construction practices and applied construction regulations. The results of research and modeling of seismic hazard and exposure are integrated in research on how to improve the calculation of seismic vulnerability of buildings. In the assessment of seismic vulnerability, vulnerability modifiers were used due to the influence of the characteristics of the foundation, for the calculation of which the spectra of uniform danger and amplification of local and deep layers of the foundation soil were taken into account. For all buildings from the exposure model and predominant typology, the values of the total vulnerability index and the proposed vulnerability classes were calculated. The difference in vulnerability index values and the variability of the vulnerability class were also investigated with respect to the most probable and less probable type of foundation soil. The results of modeling the earthquake hazard, exposure and vulnerability of the observed area we implemented in the GIS system.