The　global　food　supply　system　is　under　increasing　pressure　due　to　population　growth　and　more　extreme　climate　events.　Developing　forecast　models　for　accurate　prediction　of　crop　yields　is　helpful　for　early　warning　of　food　crise.　Amid　the　different　environmental　predictors,　soil　moisture　(SM)　information　is　an　important　agricultural　drought　indictor,　but　the　operational　microwave　SM　products　have　generally　low　spatial　resolution,　challenging　the　effective　characterization　of　the　spatial　heterogeneity　in　SM.　In　this　study,　empowered　by　the　ability　of　hourly　land　surface　temperature　(LST)　from　Geostationary　Operational　Environmental　Satellites　(GOES)　to　inform　moisture　status,　we　firstly　downscaled　SM　data　using　machine　learning　algorithms.　Then,　by　designing　three　sets　of　experiment　configurations　using　either　downscaled　SM,　coarse-resolution　SM,　or　precipitation　observation,　we　assess　the　performance　of　downscaled　SM　in　estimating　crop　yields　variability　using　three　mainstream　machine　learning　algorithms　and　two　traditional　regression　algorithms.　Our　research　shows　that　downscaled　SM　based　on　high　temporal　resolution　GOES-LST　demonstrates　outstanding　performance　in　characterizing　the　spatial　variation　of　SM.　With　respect　to　yield　estimation,　downscaled　high-resolution　SM　outperformed　coarse-resolution　SM　and　precipitation　products,　with　the　average　R2　between　the　estimated　crop　yields　and　the　yield　records　being　0.814,　0.809,　and　0.805,　respectively.　In　addition,　we　find　that　among　the　five　algorithms,　the　non-linear　machine　learning　algorithms　exceed　the　linear　algorithms,　with　the　average　R2　being　0.827　and　0.783,　respectively.　Our　research　demonstrates　the　great　potential　of　infusing　different　satellite　information　to　improve　the　monitoring　of　crop　growing　status　and　yield　prediction.　Authors