A variable selection method using global score estimation is proposed, which is applicable as a selection criterion in any multivariate method without external variables such as principal component analysis. This method selects a reasonable subset of variables so that the global scores, e.g. principal component scores, which are computed based on the selected variables, approximate the original global scores as well as possible in the context of the least squares. Three computational steps are proposed to estimate the scores according to how to satisfy the restriction that the estimated global scores are mutually uncorrelated. Three different examples are analyzed to demonstrate the performance and usefulness of the proposed method numerically, in which three steps are evaluated and the results obtained using four cost-saving selection procedures are compared.