This paper presents a new version of Davies-Bouldin index for clustering validation through the use of a new distance based on density. This new distance, called cylindrical distance, is used as a similarity measurement between the means of the clusters, in order to overcome the limitations of the Euclidean distance. The cylindrical distance takes into account the distribution of the data set, using this information to estimate the densities along line segments that connect the centroids. In this way, the index gets a more accurate measurement of separation between clusters, improving its performance.