Abstract Recent studies suggest springtime wet extremes and summertime dry extremes will occur more frequently in the US Midwest, potentially leading to devastating agricultural consequences. To understand the role of circulation patterns in the projected changes in seasonal precipitation extremes, the k-means clustering approach is applied to the large ensemble experiments of Community Earth System Model version 2 (CESM2-LE) and ensemble projections of CMIP6. We identify two key atmospheric circulation patterns that are associated with the extremely wet spring and extremely dry summer in the US Midwest. The springtime wet extremes are typically linked to baroclinic waves with a northward shift of the North American westerly jet and positive anomalies in sea level pressure over the western Atlantic, which favor the development of the Great Plains low-level jet. The summertime dry extremes are associated with the development of an anomalous ridge with suppressed storm tracks over the central US. The projected increase in springtime wet extremes and summertime dry extremes can be attributed to significantly more frequent occurrences of the associated atmospheric regimes. Particularly, the intensity of wet extremes is expected to increase mainly due to the enhanced moisture flux from the Gulf of Mexico. The moisture budget analysis suggests that the precipitation extremes are mainly associated with the dynamic component of atmospheric circulation. CESM2-LE and CMIP6 exhibit a good agreement in the projected changes in circulation patterns and precipitation extremes. Our results explain the mechanism of the projected changes in the Midwest seasonal precipitation, and highlight the contribution of circulation patterns to hydroclimatic extremes.