Subspace clustering has been successfully applied in many domains and its goal is to simultaneously detect both clusters and subspaces of the original feature space where these clusters exist. A Data stream is a massive sequences of data coming continuously. Clustering this type of data requires some restrictions in time and memory. In this paper we propose a new method named S2G-Stream based on clustering data streams and soft subspace clustering. Experiments on public datasets showed the ability of S2G-Stream to detect simultaneously the best features, subspaces and the best clustering.