ZY-1　02E　satellite　is　a　Chinese　hyperspectral　satellite　launched　in　recent　years.　There　are　few　reports　on　its　application　at　present,　so　there　is　no　suitable　research　method.　In　this　study,　rice　planting　area　was　extracted　from　two　experimental　areas　of　Fujian　Province　using　Ziyuan　No.　1　02E　hyperspectral　data.　A　multi-scale　spectral　feature　extraction　network　Based　on　spatial　neighborhood　(MSFEN-BSN)　was　proposed　to　solve　the　high　spectral　resolution,　low　spatial　resolution　and　complex　and　fragmented　distribution　of　crops　in　the　experimental　area.　The　network　consists　of　two　modules:　Multi-scale　Spectral　Feature　Extraction　(MSFE)　and　Spatial　Neighborhood　Feature　Fusion　(SNFF),　which　can　extract　spectral　features　at　multiple　scales　to　obtain　deep　spectral　semantics,　and　fuse　spectral-spatial　information　of　neighboring　pixels　to　obtain　more　effective　classification　features.　The　mapping　accuracy　of　rice　classification　in　two　experimental　areas　reached　83.61%　and　84.36%　respectively,　which　can　meet　the　needs　of　practical　application.　Meanwhile,　to　prove　the　robustness　of　the　proposed　method,　precision　verification　is　carried　out　on　two　public　vegetation　datasets　Indian　pines　(IP)　and　Salinas　Valley　(SA).　Compared　with　some　traditional　vegetation　extraction　methods　and　classical　deep　learning　methods,　the　proposed　method　has　higher　precision.　©　2024　Copyright　held　by　the　owner/author(s).