Upland　crop-rice　cropping　systems　(UCR)　facilitate　sustainable　agricultural　intensification.　Accurate　UCR　cultivation　mapping　is　needed　to　ensure　food　security,　sustainable　water　management,　and　rural　revitalization.　However,　datasets　describing　cropping　systems　are　limited　in　spatial　coverage　and　crop　types.　Mapping　UCR　is　more　challenging　than　crop　identification　and　most　existing　approaches　rely　heavily　on　accurate　phenology　calendars　and　representative　training　samples,　which　limits　its　applications　over　large　regions.　We　describe　a　novel　algorithm　(RRSS)　for　automatic　mapping　of　upland　crop–rice　cropping　systems　using　Sentinel-1　Synthetic　Aperture　Radar　(SAR)　and　Sentinel-2　Multispectral　Instrument　(MSI)　data.　One　indicator,　the　VV　backscatter　range,　was　proposed　to　discriminate　UCR　and　another　two　indicators　were　designed　by　coupling　greenness　and　pigment　indices　to　further　discriminate　tobacco　or　oilseed　UCR.　The　RRSS　algorithm　was　applied　to　South　China　characterized　by　complex　smallholder　rice　cropping　systems　and　diverse　topographic　conditions.　This　study　developed　10-m　UCR　maps　of　a　major　rice　bowl　in　South　China,　the　Xiang-Gan-Min　(XGM)　region.　The　performance　of　the　RRSS　algorithm　was　validated　based　on　5197　ground-truth　reference　sites,　with　an　overall　accuracy　of　91.92%.　There　were　7348　km2　areas　of　UCR,　roughly　one-half　of　them　located　in　plains.　The　UCR　was　represented　mainly　by　oilseed-UCR　and　tobacco-UCR,　which　contributed　respectively　69%　and　15%　of　UCR　area.　UCR　patterns　accounted　for　only　one-tenth　of　rice　production,　which　can　be　tripled　by　intensification　from　single　rice　cropping.　Application　to　complex　and　fragmented　subtropical　regions　suggested　the　spatiotemporal　robustness　of　the　RRSS　algorithm,　which　could　be　further　applied　to　generate　10-m　UCR　datasets　for　application　at　national　or　global　scales.　©　2024　Crop　Science　Society　of　China　and　Institute　of　Crop　Science,　CAAS