Abstract: A method for quantitative reconstruction of paleowater depth based on Milankovitch cycles is provided, which comprises following steps: selecting a calibration well rock samples lithofacies sensitive logging data, performing major and trace element analysis, and calculating a single-point paleowater depth; denoising the lithofacies sensitive logging data; performing bandpass filter analysis and multitaper method (MTM) spectral analysis on the denoised well logging data to determine the applicability of the denoised data, followed by performing Evolutive Harmonic Analysis (Eha) & Evolutive Power Spectral Analysis and calculating Evolutive Average Spectral Misfit (eAsm) by Monte Carlo simulation, thereby obtaining a depth-domain spectrum; based on the Milankovitch astronomical cycles theory, tracking a minimum of a null hypothesis significance level; establishing an equal-depth correspondence between the obtained sedimentation rate and the single-point paleowater depth, fitting a sedimentation rate-paleowat