As the transfer function of the near field acoustic holography (NAH) algorithm, the accurate acquisition of Green Function Fourier spectrum is the prerequisite for the high precision reconstruction of the sound field. In this paper, the basic principle of NAH is presented and two kinds of the acquisition methods of the Green Function’s Fourier spectrum are introduced in detail. Then through the simulation of underwater sound source for the reverse reconstruction so as to complete the sound source localization. The results show that when the sound source is less, the positioning error of the two methods were less than 5 cm, and the acoustic pressure amplitude of the source is proportional; while when the sound source number is larger, the sound field reconstruction based on the Kspace sampling Green function has the robustness to suppress virtual sound sources generated by the multiple sound sources, which is more suitable for the location of multiple noise sources.