Horel and Kennard gave the generalized ridge regression(GRR) in 1970.It is a new multivariate analysis method. In recent years it has developed at theory and application rapidly. Same as ridge regression and partial Least-squares regression, GRR mainly applies to multicollinearity and is effective settlement for sample measure less than the number of independent variable. It is difficult to compute inverse matrix and select ridges.In particular, chemistry and biologic tests often appear the number of unknown parameters p far more than sample measure. This moment matrix Xn×ppossess multicolinearity. By reason of the test or process breakup or outlay restricted, etc, n less than 100, probably much less. This produces an especial data matrix. In view of Xn×p(n 2) speedy iterative arithmetic of GRR as to Xn×p(n |