Subspace　clustering,　known　for　its　effectiveness　in　handling　high-dimensional　data,　has　attracted　attention.　And　the　autoencoder　can　discover　hidden　features　within　a　large　dataset,　yet　it　faces　challenges　in　utilizing　attribute　information　for　reconstruction　and　capturing　complex　spatial　structural　information.　To　tackle　these　issues,　we　propose　a　community　detection　algorithm　named　Deep　Attention　Autoencoder　Based　on　Subspace　Constraints　(DAASC).　First,　we　design　an　attribute　topology　fusion　strategy　to　integrate　attribute　information　into　the　reconstruction　of　the　decoder.　Then,　we　design　a　subspace　autoencoder　strategy,　using　the　concept　of　subspaces　to　construct　the　loss　function,　to　capture　the　spatial　structural　information　of　the　data.　Experiments　conducted　on　both　real-world　and　synthetic　networks　to　compare　DAASC　with　several　state-of-the-art　community　detection　algorithms　demonstrate　its　exceptional　accuracy　and　robustness.　©　2024　IEEE.