As　a　new　network　paradigm,　Software-defined　networking　(SDN)　realizes　centralized　management　of　the　network　by　separating　the　control　plane　and　the　data　plane.　While　SDN　greatly　improves　network　management　capabilities,　it　also　brings　some　security　risks　such　as　Distributed　Denial　of　Service　(DDoS)　attack.　How　to　effectively　detect　abnormal　traffic　has　always　been　a　hot　issue　in　the　field　of　network　security.　This　paper　proposes　an　improved　attack　detection　model　SSAE-BiLSTM　based　on　deep　learning.　The　stacked　sparse　autoencoder　(SSAE)　is　used　to　extract　high-dimensional　features　of　data,　and　bidirectional　long　short-term　memory　(BiLSTM)　is　used　to　classify　network　traffic.　This　model　can　effectively　detect　network　attacks　with　higher　accuracy　and　lower　false　alarm　rate　on　the　benchmark　dataset　UNSW-NB15.　©　2021　IEEE.