Abstract: Permeability coefficient is not only an important index for describing the soil permeable property but also an important factor to influence the consolidation settlement and seepage deformation.Firstly,we have used 28 sand samples in the Middle East of the North China Plain to do consolidation tests and permeability tests by a self modified osmotic oedometer and then have obtained amounts of permeability coefficients data under different void ratios.Then,we have used principal component analysis method to extract 3 principal components,which is related to effective diameter d10,limited diameter d60,non-uniform coefficient Cu,curvature coefficient Cc and void ratio e,as factors of permeability coefficient and have explained the principal components from different angles.Finally, we have introduced the support vector machine method to establish the permeability coefficient prediction model of sand soil in the Middle East of the North China Plain and randomly choose 20 sand samples in this area to predict.The results show that the osmotic oedometer can be better used to conduct consolidation tests and permeability tests;the accuracy of this permeability coefficient prediction model is relatively high which can provide reference for the research of sand soil permeability coefficient in this area.