Abstract: Obtaining reservoir rock mechanics data is the basic work for sanding prediction. The most reliable way to obtain the core mechanics data is through the experimental measurement of the core, but the rock mechanics experiments often have certain deviations and limitations. In order to obtain effective rock mechanics parameters, this paper uses more easily obtained logging data to find the internal relationship between logging data and experimental data, and then predicts the rock mechanics parameters. In this paper, the statistical analysis of the uniaxial compressive strength experimental data of dozens of cores in 4 oil fields in Bohai Oilfield shows that the uniaxial compressive strength is strongly correlated with logging data such as acoustic time difference, depth and density. Using multiple generalized linear regression, the rock mechanics parameter prediction formula is established. The rock data of the newly developed oil field were used to verify. The results show that the new formula error is much smaller than the conventional empirical formula, and the absolute error value is within 1.5 MPa. The formula has good adaptability and reliability in Bohai Oilfield.