With the depletion of traditional energy sources, the usage of renewable energy alternatives has increased over the past years. Among them, wind energy is considered as one of the most economical and environmentally friendly alternatives, and is seen as one of the main sources of energy in the future. A very promising field in the area of wind energy is offshore wind energy. Its currently high costs in comparison with onshore wind energy are forecasted to decrease in the coming years. Moreover, an acceleration in the deployment of the offshore wind industry is expected not only in Europe, but also in other regions such as Asia and America. With such development, which is also linked to an increase in the capacity of the turbines, one topic that has only been recently addressed is the effect of earthquakes on offshore wind turbines (OWT). There exists no clear guideline on how a seismic analysis should be performed in the design of OWTs as their dynamic behavior is far more complex when compared to regular buildings. Several authors have studied this topic and have found that aspects such as the aerodynamic damping of the turbine, the higher order mode influence or the soil-structure interaction should be taken into account, as they have a significant influence on the response of the turbine due to earthquake loading. However, there is a general lack of knowledge on how to include these aspects in the analysis, especially in combination with regular OWT design loads. Moreover, it is still unknown if earthquake loading could actually become design-driving for OWT design, which is currently governed by wind and wave load cases. This thesis aims to fill this knowledge gap by addressing all the aforementioned questions, up to a certain extent. As a first approach into getting an insight on the effect of seismic loading on OWT design, a linear finite element model of an OWT monopile was developed. This model accounts for the soil-structure interaction by using a set of distributed linear springs and dashpots to model the soil stiffness and damping, respectively. Moreover, these springs and dashpots are used to apply a set of ground motions (as displacements and velocities), which were selected and matched to a target response spectrum, and were subsequently calculated for different depths by using shear wave one-dimensional ground response analysis. The developed embedded monopile model was coupled with 4.0 MW wind turbine model using BHawC; the existing Siemens Wind Power in-house code for wind turbine nonlinear analysis in the time domain. This code combines elasticity and aerodynamics in order to calculate the loads induced on the turbine components due to different load cases, while accounting for its different operational states and the changes on the turbine dynamic properties because of them. Moreover, it also models features and systems that are present in the actual Siemens wind turbines to optimize its power production while minimizing the structural damages. Initially, a flexible soil-pile model was assumed along with relatively stiff soil conditions. It was found that the response is dominated by higher order eigenmodes during the earthquake event, and first order tower bending modes after it has passed. Moreover, a large influence of blade modes was observed along the whole height of the support structure, but particularly at the upper part (tower top and blades), where significant differences in the response were noted between both tower directions (fore-aft and side-side) and different modeled operational states. When comparing the results with the actual turbine design loads, it was found that the tower top and the blades (in the edgewise direction) could become design-driving if earthquake loading is present. Due to the influence of the soil properties in the overall system, two additional soil cases (medium and soft) were subsequently considered. It was found that the soft soil led to the highest load due to an amplification of the response in the frequency range in which the turbine was mostly excited. The effect of the ground response analysis could also be observed, as the response was further amplified for frequencies close to the fundamental frequencies of the soil deposit. Furthermore, new analyses were performed with variations in the model such as the soil damping, the soil-pile model and the ground response model. It was found that the way the soil is modeled can have a large impact in the design loads, particularly regarding the choice modeling approach taken for the soil-structure interaction (the spring and dashpot coefficients), albeit the choice of the dashpot coefficients alone did not have a significant influence on the results. Moreover, it was found that modeling the monopile as a single discrete spring and dashpot system through its equivalent dynamic impedances can lead to an important underestimation of the loads, which can be compensated by applying the actual monopile displacements and rotations induced by the earthquake loading, to account for the kinematic soil-structure interaction. As a main outcome of this study, it was concluded that earthquake loading can in fact become design-driving, particularly at the upper portion of the turbine due to the influence of the higher modes. In this context, it was highlighted that aspects such as the blades and the different turbine operational states should be included in the analysis, as they have a significant influence on the results. The importance of an accurate soil identification and modeling, along with a proper choice of ground motion time series were emphasized. Finally, recommendations for future research were given , particularly regarding the inclusion of aspects such as non-local soil-pile interfaces, soil nonlinearities, wave loading and frequency domain methods in the analysis.