To meet with the need of extracting clustering boundary from mixed attribute data set in field of data analysis, a clustering boundary detection algorithm for the mixed attribute data set is proposed, named BERGE. BERGE defines the boundary factor based on the membership of fuzzy clustering to recognition the candidate boundary set, and utilizes the idea of evidence accumulation to extract clustering boundary from the candidate boundary set. The experimental results on synthetic data sets and real data sets show that BERGE can effectively detect clustering boundary of mixed attribute data sets, numerical attribute data sets and categorical attribute data sets with higher accuracy compared with the existing similar algorithms.