Abstract: Prediction of sand liquefaction trend is an important basis for sand liquefaction control. In this paper,a new combination weighting model(AHP-RS) was proposed based on the analytic hierarchy process(AHP)and rough set theory model(RS),combined with the distance function. The AHP-RS model was used to evaluate the liquefaction trend of sandy soil in the southern Tangshan earthquake disturbance area, and the prediction results of the three evaluation models of AHP,RS,and AHP-RS were compared with the actual liquefaction. The results showed that:(1)According to AHP-RS model, mild and medium liquefaction areas accounted for 49.9% and heavy liquefaction area accounted for 50.1%, proving that the results accuracy by this method was greatly improved compared with those by AHP and RS.(2) From perspective of the grades classification, the division results of the AHP-RS model were closer to the actual liquefaction distribution, confirming that the prediction results of the sand liquefaction evaluation by the AHP-RS model had higher accuracy and applicability. The new model meets actual engineering needs and can be further promoted in application.