Abstract: Seismic exploration in complex areas is often constrained by challenging terrains such as mountains, farmlands, and residential zones, making regularized acquisition difficult. Furthermore, in merged-survey processing, discrepancies in acquisition geometries across different areas often hinder seamless data integration at the junction zones. Irregularly sampled seismic data tends to suffer from spatial aliasing and acquisition footprint artifacts during processing. Five-dimensional (5D) regularization is required to regularize and homogenize the data distribution, thereby enhancing the quality of subsequent modeling, imaging, and interpretation. Existing regularization methods primarily employ the Anti-leakage Fourier Transform (ALFT) to compute the spectrum of irregularly sampled seismic data, followed by inverse transformation to reconstruct the data onto predefined coordinates. However, for 5D seismic data, the exponential increase in data volume leads to prohibitive computational costs for Non-uniform Discrete Fourier Transform (NDFT) calculations. Combined with the hundreds of iterations required by ALFT, 5D regularization becomes unfeasible for industrial application. Therefore, this study proposes a fast seismic data regularization using double lookup tables method for seismic data. By pre-computing natural exponential terms, the computational burden of NDFT is significantly reduced; meanwhile, replacing iterative NDFT calculations with table lookup operations requires only a single forward-inverse transform pair to achieve results comparable to conventional ALFT, achieving approximately 100 times higher computational efficiency. Currently, this method has been successfully applied to multiple Sinopec surveys, with processing results from various field datasets demonstrating its effectiveness.