Résumé : Early in an offshore wind farm development, the project developer needs to determine which type of offshore wind foundations are suitable to its site. At that stage, only a limited amount of site-specific data is available. However this conceptual study is critical to estimate a preliminary cost target and assess the techno-economic viability of an offshore wind farm at that location.In this context, INNOSEA has developed a user-friendly preliminary design tool for foundations (monopile, jacket and gravity-based foundation) named PREDIN. This numerical software tool is able to perform semi-automatically: the conceptual design, the planning of the installation phase, the cost estimate of fabrication, transport and installation of foundations and substations for sound decision-making. The software features a lean graphical user interface which provides an integrated vision: from foundation design to planning and key cost drivers. PREDIN is adapted for uninitiated users, by providing guidelines and standards values, and for specialist users by enabling a total control of the process.The foundation design methodology is based on a two-step approach: General conception, mainly based on verification of the dynamics of the structure; Local sizing: PREDIN interfaces with a standard structural software from the offshore industry: ANSYS Structural. This two-step approach therefore makes it possible to be able to rapidly iterate on the foundation design.The transport and installation module methodology relies on 4 different functions: Weather Data extraction, Marine Spread selection, Operabilities and T&I planning.The cost estimate methodology is based on a Monte Carlo probabilistic approach, integrating the projects uncertainties in order to generate the probability distribution of the final cost.This study highlights the capabilities of PREDIN software to efficiently deliver a preliminary design and EPCI cost of offshore wind foundations for project developers, relying on extensive metocean, turbine, foundation and cost databases.