To design well paths that minimize risk and avoid drilling hazards such as blowouts and stuck pipe, drilling engineers would like to have a quantitative understanding of the overpressure zones in the subsurface. Currently, pre-drill prediction of pore pressure is done using kinematically determined seismic velocity, which has a low resolving power in identifying various subsurface formations. In the rare examples where high-resolution velocity is used, the primary seismic input is inverted acoustic impedance. The acoustic impedance is converted into high-frequency velocity and density for effective stress and overburden stress computations. Both require transformation schemes, potentially causing additional uncertainty in pore-pressure prediction. In this paper, we present a method based directly on acoustic impedance. We thus avoid the additional, potentially error-prone step of converting impedance to velocity and density. We modify the methodology described in Rasolofosaon and Tonellot (2011). We call this the RT method in this paper. Using well log data, we first demonstrate that the RT method provides practically the same results as those using velocity and density data at the well location, and does it more efficiently. This leads us to suggest that the formation pore pressure itself can be written as a piece-wise continuous function of a single variable, acoustic impedance. This greatly simplifies the work steps in pore-pressure prediction methodology. This new method is then applied to well and seismic data in deepwater Gulf of Mexico (GoM) subsalt basins, predicting subsalt and salt-exit pore pressure. We compare the predicted results with measured pore-pressure data where available.