Abstract: In a quantitative drug design, correlation coefficient matrix of physico-chemical indices X_i is often pathological when there indices are highy correlated. Theregression equation of biological activity Y about there indices X_i obtained in this case isnot stable. In this paper, a method is given to obtain a stable regression equation:giving acritical valne α and finding out the indices among which correlation coefficient is not lessthen α, The following are the rules to reject some of them. If X_i and X_j is highly correlated(|r_ij|≥α, r_ij is the correlation coefficient of X_i and X_j, and |r_in|>|r_jn|(r_in and r_jn are correlation coefficients of X_i and X_j about Y, respectively), then the index X_j is rejected, other-wise, X_i is rejected. Stable equation can be obtained by stepwise regression with the re-maining indices.