Abstract: Wet collapsibility coefficient is a common index to judge the class and risk of loess collapse, and to predict the volume of wet collapse. In order to get more conveniently and accurately the wet collapse coefficient of loess in engineering construction site, soil samples have been taken in Qishan and Xinping of Shaanxi Province to carry out geotechnical test in laboratory and to get different physical parameters. Through statistical analysis, it was found that there is good relation among loess collapsibility coefficient, water content, dry density, void ratio and compressive modulus. A series of regression equations have been got through quaternion, ternary and binary fitting of curves. Through simulation and prediction using regression equations it has been found that it is feasible to predict the collapsibility of loess in Guanzhong area by regression equations.