In this paper, we improve the classification performance of categorical data using an Adoptive Hamming Distance. We defined the equivalent categorical values and showed how those categorical values were searched to adopt the distance. The effectiveness of the proposed method was demonstrated using various classification examples.