Introduction Understanding the spatial variability of crop yields is important for improving agricultural resilience in semi-arid regions where limited water resources and the growing impacts of climate change present significant challenges to agricultural sustainability.Methods This study applied the DSSAT model to assess the spatial and temporal variability of cotton, sorghum, and winter wheat yields in response to climate and soil variability in the Texas High Plains. The predictions were validated using USDA-NASS county-level annual yield data and the Cropland Data Layer (CDL).Results The model showed good performance in simulating crop yield, with average R2 values of 0.57 for cotton, 0.72 for sorghum, and 0.71 for wheat, alongside normalized root mean square error values of 14.02%, 11.15%, and 15.25%, respectively. However, model performance varied across counties and growing seasons, with underprediction of cotton and sorghum yields in 2012 following the severe 2011 drought. Spatiotemporal analysis from 1980 to 2022 revealed distinct patterns of yield variability across soil types and crops. Cotton and wheat showed slight increases in yield over time on finer-textured soils such as clay loam, silty clay loam, silt loam, and loam, while sorghum yields remained relatively stable across these soils. In contrast, all three crops exhibited slightly declining yield trends on sandier soils, with cotton showing the greatest variability, followed by sorghum and wheat.Discussion These results underscore the importance of accounting for soil type when assessing the impacts of long-term climate variability on crop yield. The observed yield patterns offer valuable insights for site-specific crop management and climate adaptation strategies in the region.