The seismic assessment of existing buildings has been underestimated in regions with moderate seismic hazard, such as the North-West Europe. Strong earthquakes, such as the 1356 Basel earthquake, already occurred in such regions and caused severe damage. The majority of the building stocks in those regions were built before the introduction of the modern building codes. Unreinforced masonry structures are widely used, and post-earthquake surveys show that such kind of structures are vulnerable under seismic actions. Moreover, existing reinforced concrete structures, which do not integrate current seismic features, are also highly vulnerable. Fragility analysis, which permits to estimate the probability (of an individual building or a building class) of reaching or exceeding a damage grade under a given earthquake load level, arises as a useful tool to evaluate seismic risk in existing buildings. Although several methodologies to derive fragility curves already exist, they consider building typologies which are not necessarily representative of regions with a moderate seismic hazard. The methodology presented in this dissertation allows engineers to derive fragility curves for unreinforced masonry and reinforced concrete structures, using a simplified and analytical methodology. Therefore, analyses for building classes or preliminary results for individual buildings can be obtained. The procedure is based on an improved static non-linear approach, and on analytical definitions of damage grades. In general, existing buildings can develop a moderate torsional behaviour; hence, the methodology integrates that phenomenon in terms of damage grades. With those damage grades, and considering the epistemic and aleatory uncertainties inherent in the analysis, fragility curves can be derived. Given that spectral parameters are better descriptors of earthquakes than other ground motion parameters, fragility curves are expressed in terms of spectral displacement. The methodology is capable to derive damage estimations in a reasonable amount of time, compared to more detailed analytical approaches. The results obtained in terms of capacity curves, modal parameters and probability of damage (an application of fragility curves) show a satisfactory correlation with experimental data. Finally, several real buildings are studied and the corresponding fragility curves and probability of damage under a seismic scenario are presented and compared to those of existing methodologies.