<p>Subspace clustering, particularly multi-view subspace clustering, has become increasingly relevant in machine learning due to the proliferation of multi-view data sets. Despite significant advancements, existing multi-view subspace clustering algorithms still encounter two primary limitations. Firstly, most methods learn the affinity matrix using constraints or regularization terms and then apply spectral clustering. This approach is susceptible to noise and redundant information in the original data. Secondly, in practical applications, multi-view data is often stored across different devices, some of which may contain private information that cannot be shared. Although traditional federated learning can address this issue, this approach faces limitations in scenarios where the central server is either absent or has failed. To resolve these problems, we propose a Decentralized Federated Multi-view sparse Subspace Clustering(DFMSC) method. DFMSC introduce a decentralized approach that avoids the need for a central server, reducing the vulnerability to server failure and enhancing data privacy. Specifically, our approach integrate self-representation learning, graph structure updating, and spectral embedding learning within a decentralized framework. We enforce consistency across different views by introducing a consistency constraint, which ensures that updates are made locally while achieving a unified spectral embedding through neighbor communication. Accordingly, we propose an iterative algorithm to solve the resulting optimization problem. Experimental results on a variety of real-world multi-view datasets demonstrate the superiority of our approach.</p>