Abstract: Void is the main factor leading to the fracture of cement slab, the early location of void area can provide a basis for the slab structural safety analysis. Void area in the cement pavement was taken as the target, and its influence factors, including void shape and filling medium, were considered and analyzed with FDTD method. CWT (continuous wavelet transforms) method was adopted to process the GPR signal. To obtain the original parameters of GPR signal, a GPR signal processing module was designed. Our proposed void feature extraction method was tested both in simulated and lab model. The results indicate that the designed software can realize the post-processing function of GPR signal, the normal and void area have significant difference in time-frequency domain after CWT process. The energy concentration area of CWT map can represent the depth and range of void area, which could be also viewed in the CWT 3D image and used to determine the void area whether contain water or not. Our method provides a foundation for GPR signal processing.