Abstract: Based on the theory of maximum and minimum void ratios of granular materials,a new liquefaction potential index was used to predict the liquefaction resistance of binary sand silt mixtures with different particle gradations. By calculating the volume strain potential of sand silt mixtures in the existing literature data,the relationship among the fine particle content,the particle size ratio and the cyclic resistance ratio CRR of the mixture was analyzed. A new liquefaction potential index \varepsilon_\mathrmv^\prime was proposed combined with the actual void ratio of the mixture. The comparative analysis with the CRR test data of sand silt mixtures shows that the CRR of all kinds of sand silt mixtures decreases monotonously with the increase of \varepsilon_\mathrmv^\prime,and the two parameters show a good power function relationship. Further,the relationship was established for the best fitting parameters of CRR-\varepsilon_\mathrmv^\prime relationship curve and particle size ratio. The CRR prediction results of Monterey sand mixed Yatesville silt samples show that the \varepsilon_\mathrmv^\prime comprehensively reflects the comprehensive effects of soil particle size distribution,compaction state and particle shape,and is an effective index to characterize the CRR of sand silt mixtures. In practical engineering,it can replace the traditional in-situ test and predict the anti-liquefaction capacity of site soil layer only through simple indoor test.