In this paper, the linear regression model based on the financial indicators of listed real estate companies and their stock price was established. The least squares regression, principal component regression and stepwise regression methods were used to estimate the regression coefficients. 〖JP2〗Compared with these three methods, variable selection based on the stepwise regression method was the most comprehensive and reliable. Furthermore, multivariate linear regression equation between the index of the real estate companies and their component stocks was established, and we used the elastic restricted estimation method to implement variable selection of component stocks. Meanwhile,both problems of unsignificance of the regression coefficients and stock index tracking were also solved.