One-way analysis of covariance is a popular and common statistical method, wherein the equality of the means of several random variables which have a linear relationship with a random mathematical variable, is tested. In this study by using Dlr metric, a method is presented to improve the one-way analysis of covariance when there is an uncertainty in accepting the statistical hypotheses. The method deals with a fuzzy test statistic which is produced by a set of confidence intervals. Finally an example is provided for illustration.