Touchless fingerprint capturing can provide richer features and overcome the deficiency of 2D fingerprint recognition. In this paper, a clustering-based dynamic score selection (CDSS) algorithm is proposed for the combination of scores which are generated by different vision touchless fingerprint recognition systems. First, the scores are divided into two classes and the number of elements in each class and other statistic variables is computed. Then appropriate statistic value is chosen as the score for final decision of the whole system. The experimental results show that the performance of CDSS-based multi-vision touchless system can be enhanced efficiently compared to touchless fingerprint recognition and better than those of sum, max, SVM and Fisher linear discrimination algorithms.