Subsalt Gulf of Mexico deep-water wells routinely cost in excess of $100 million. A reliable pore pressure prediction can translate into considerable savings in terms of drilling costs and safety. Traditional methods used to determine pore pressure are based on either logs (e.g., Eaton's or Bowers' methods) or seismic data (e.g., calibrated seismic velocities, acoustic impedance). Another method for pore pressure prediction is based on basin modeling: building a three-dimensional earth model and simulating the processes of pressure formation, through geologic time. Recent advancements in basin modeling, such as the coupling of stress and pressure and the implementation of models for mineral diagenesis and rock failure, have significantly improved its applicability. However, no single method is commonly accepted as better than another, therefore, using, comparing, and integrating all three methods together in a predrilling project can provide a higher degree of confidence for pore pressure prediction. The purpose of this paper is to describe a new approach to pore pressure prediction that combines the above methods with petroleum system modeling. A special emphasis is put on the explanation of the basin modeling workflow. The first step of the workflow is to create and calibrate a regional model based on a set of regional maps with the main goal of providing the regional context. The second step is to create a smaller area of interest (AOI) model using high-resolution structural and facies maps. This refined model is then used for pore pressure prediction at the prospect scale. The smaller AOI model, albeit at very high resolution, allows a model to be run overnight, so that pore pressure can be predicted ahead of the drilling bit. Finally, the predicted pore pressure and fracture gradient allow the drilling engineer to optimize well performance and reduce drilling costs.