With　the　development　of　big　data　artificial　intelligence,　compressed　sensing　spare　reconstruction　(CSSR)　problem　has　always　been　a　research　hotspot　in　the　field　of　signal　processing.　To　deal　with　the　problem,　the　CSSR　problem　is　first　viewed　as　a　multi-objective　optimization　algorithm.　And　a　multi-objective　CSSR　model　is　explained　by　considering　the　factor　of　reconstruction　error　and　sparsity　degree.　To　solve　the　optimization　model,　an　immune　algorithm　based　on　neighborhood　selection　is　employed.　And　the　value　of　reconstruction　error　and　sparsity　degree　could　be　optimized　by　using　the　nondominated　neighbor　immune　algorithm　method　to　obtain　the　minimize　value.　Meanwhile,　to　verify　the　performance　of　sparse　reconstruction　four　one-dimensional　sparse　signals　are　adopted.　The　simulation　result　show　that　the　reconstruction　error　of　the　nondominated　neighbor　immune　algorithm　obtains　the　best　result　on　the　whole　and　the　performance　improve　rate　of　can　reach　40.00%　at　most　on　sparse　signal　F1-F4.　©　2021　ACM.