Accurate prediction of pore pressure can assist engineers to better work out and optimize an oilfield development plan. Because the P-wave velocity only experiences small-scale fluctuations for pore-pressure change in carbonate rocks, existing well-known pore-pressure prediction methods are incapable of predicting pore pressure in carbonate rocks with field-required accuracy. We evaluated a new method based on the P-wave velocity decomposition and wavelet transformation to predict pore pressure in carbonate rocks. The P-wave velocity was decomposed into contributions from the pore fluid and the rock framework using Biot’s theory. The effect of lithology, pore structure, porosity, and pore pressure on P-wave velocity was studied by theoretical analysis and experiments. Rapid triaxial rock-system tests were carried out to measure the P- and S-wave velocities when pore pressure, pore structure, and porosity were changed, and X-ray diffraction tests were used to measure mineral components. The small-scale fluctuations of the P-wave velocity can be extracted and amplified using wavelet transformation. We found that the small-scale fluctuations of the P-wave velocity were caused by pore-pressure change in carbonate rocks and the large-scale fluctuations of the P-wave velocity depended on the rock framework. Overpressure formation can be identified by the high-frequency detail of wavelet transformation of P-wave velocity. A pore-pressure prediction model relating the contribution from the pore fluid to the P-wave velocity was developed. This model is an improvement over existing pore-pressure prediction methods that mainly rely on empirical relations between the P-wave velocity and the pore pressure. This new method was successfully applied to carbonate rocks in Tazhong Block, Tarim oilfield, demonstrating the feasibility of the proposed pore-pressure prediction method.