Abstract: The cone penetration test (CPT) is a commonly used technique to identify the liquefaction of sandy soil in the world but it cannot reflect the influences of pore water pressure (pore pressure) on the liquefaction behavior. The piezocone penetration test (CPTU), a modern in-situ test technique, can measure pore pressure, and has high identification and sensitivity to sandy soil and silty soil, so it has a unique advantage in liquefaction evaluation of sandy soil and silty soil. In this study, a liquefaction resistance ratio (CRR_7.5) model based on CPTU test parameters is established under the simplified stress framework through the particle swarm optimization (PSO)-kernel limit learning machine (KLEM) algorithm, robust search technology and nonlinear fitting optimization analysis. The model includes modified cone resistance (q_t1N) and soil behavior type index (I_{c, BJ}) as per Jefferies and Davies, including the pore pressure ratio (B_q). In this way, the soil type information in the proposed model is directly included in the liquefaction discrimination formula, and it does not need to modify the penetration resistance according to the content of soil fine content, which is more direct and more consistent with the mechanical framework, and is suitable for a wide range of soil, such as sandy soil and silty soil. The liquefaction discrimination of sandy soil and silty soil is unified by the I_{c, BJ} with the inclusion of B_q in the formulation. Finally, the accuracy and superiority of the proposed model are proved by a case study of Tangshan Earthquake liquefaction in China.