Abstract: A new hyperbolic model for estimating liquefaction potential of sand is proposed to solve the exiting problems in the available liquefaction discrimination methods. By collecting 156 in-situ SPT data from previously liquefied and non-liquefied sites in China, a new SPT-based formula is proposed. The reliability of the new formula is verified by 312 data collected from 1995 Kobe Earthquake and 1999 Chi-Chi Earthquake. The estimated results by the new formula are compared with those by the code method and Seed method, indicating that the hyperbolic model is quite adaptable for a wide range of seismic intensities, ground water level and buried depth of sand. The new model compensates the limitation that the discriminated results of shallow sand (less than 10 m deep) tend to be risky by the code method under intensity Ⅶ. Using the new model, the success discrimination ratios of liquefied and non-liquefied sites for shallow sand under intensities Ⅷ and Ⅸ are more uniform than those by the code method. The critical discriminating curve of the new model meets the fact that its slope is large in shallow while small in deep one. The disadvantage that the discrimination of soils 10 m to 20 m in depth under intensities Ⅷ and Ⅸ is significantly conservative by the code method is corrected. The proposed model adopts asymptotic form which is closely consistent with the real condition and avoids the abnormal phenomena using the Seed method in which N_cr increases and afterward decreases with depth.