In the past decades, the wind industry has grown from a niche business serving the environmental aware into one that has established itself as the most competitive form of renewable energy. Wind has the potential to play a more important role in the future world electricity supply provided that the cost per kilowatt-hour are further reduced. The cost of wind-generated electricity can be effectively reduced by steady improvements in both wind turbine design and operation since the conversion of wind energy into electricity is a highly capital-intensive and maintenance-demanding technology. The challenge of wind energy research lies in developing wind turbines that are optimized with respect to both cost and performance. A prerequisite for the cost- effective design of such turbines is the availability of a systematic methodology that generates accurate and reliable dynamic models of the complete system within the design phase with relatively low modeling effort. The models, in turn, can be used to evaluate the impact that the design choices have on the economic viability. Sub- sequently, the dynamic behavior of the selected wind turbine configuration can be assessed under various (extreme) conditions and, when required, improved. In this thesis such a methodology has been developed for the class of wind tur- bines that has the highest potential to reach cost-effectiveness in the (near) future. The validity of the methodology has been judged by means of experiments per- formed on a direct-drive synchronous generator, various rotor blades, and a full- scale wind turbine. In addition to this, the first step is made in using the developed models for control design. The developed methodology has been implemented in a user-friendly design code called DAWIDUM. DAWIDUM has been developed in the MATLAB R /SIMULINK R environment. It can be concluded that in this thesis the basis has been laid for the cost-effective design and operation of variable speed wind turbines.