To meet growing energy demands, the Kyoto protocol and the much desired diversification of supply, wind energy has become a mainstream source of energy in the EU. Cost wise it is already competing with gas fired electricity. In the last decade wind moved offshore to accommodate even more wind power. The offshore wind resource is more abundant and of a better quality, resulting in higher electricity output. On the other hand, the cost of installing turbines offshore is higher than onshore. To improve the cost-effectiveness of offshore wind, the risks involved must be known and mitigated and the critical design parameters must be optimised. From an engineering point of view, these requirements can be met through the following steps: - understand the basics of offshore wind turbines - apply lessons learned from previous projects - improve design tools. This thesis focuses on the design of the support structure. First, the basics of offshore engineering and of wind energy technology are summarized, specifically focused on the support structure design. Then, an overview is given of four actual offshore wind farm designs and their details. The design methods were compared mutually and with a design of a typical offshore oil platform. For most of the design steps, the methodology is consistent. Only the fatigue damage assessment is done differently for each individual project. Fatigue assessment in offshore engineering is done in the frequency domain. This method can be applied because the wave loads can be effectively linearized. The advantages of the frequency domain method are the clarity of presentation of intermediate results and the final outcome as well as the speed of calculation. The offshore wind industry standard (both onshore and offshore) is to use time domain simulations, which enables taking all non-linearities of the turbine operation into account. A disadvantage of this for the design of support structures is that offshore contractors lack both the aerodynamic knowledge and knowledge of the turbine details to use the full time domain simulation method to calculate the total fatigue damage. In this thesis a frequency domain method is developed to solve this problem. An interface between turbine manufacturer and offshore contractor is created that avoids the need to transfer commercially sensitive turbine details. The offshore contractor can further optimise the support structures with the software packages he normally uses. The frequency domain method is tested for the Blyth offshore wind turbines, for which a validated computer model and on-site measurements were available. Further, the method is applied to a design for the Dutch offshore wind farm to be erected at Egmond in 2006. In both cases, the frequency domain method works very well and gives results that compare well with time domain results. The computer time required to perform a fatigue calculation has been reduced from several hours in the time domain to less than 2 minutes in the frequency domain. This high speed of calculation opens possibilities for parameter variations to check the sensitivity of design choices and for optimisation of every structure within the wind farm. This has the potential to significantly reduce cost and risk. A key issue in the accuracy of the method is the effect of the aerodynamic damping of the operating turbine on support structure dynamics. Several calculation methods for this damping have been tested and have shown to give reasonable results. More work is needed to more accurately pinpoint the magnitude of this aerodynamic damping. The frequency domain method is currently being implemented in the software of an offshore contractor while other companies have already shown interest.