The double bass, as the instrument with the lowest timbre and the largest volume in the string section of a symphony orchestra, is the “mainstay” of the orchestra’s acoustic effect, and grasping the bass performance mode in double bass performance is a problem that all double bass players need to explore in depth. A cluster-weighted multi-view kernel k-means clustering model (CWK2M) is proposed to study the local quality differences of the bass performance score views at the cluster level. The proposed weighted multiview clustering algorithm is then compared with several multiview clustering algorithms on several real multiview data for experiments and analysis of pitch change patterns. The experimental results show that, on the whole, the proposed algorithm in this paper obtains a relatively good clustering effect on each multiview data, especially on the Sens IT dataset of bass performance scores, the performance of each metrics is significantly improved, and the precision, recall, F1 value and NMI metrics are 0.632, 0.653, 0.687, and 0.713, respectively.In addition, the algorithm of this paper is utilized for the three bass playing patterns such as TaS1, Py11 and Mla1 are further analyzed, which further validates the universality and performance effect of the improved weighted clustering algorithm proposed in this paper for the analysis of pitch change patterns in bass playing.