Restoration in this era of climate change comes with a new challenge: anticipating how best to restore populations to persist under future climate conditions. Specifically, it remains unknown whether locally-adapted or warm-adapted seeds best promote native plant community restoration in the warmer conditions predicted in the future and whether local or warm-adapted soil microbial communities could mitigate plant responses to warming. This may be especially relevant for biomes spanning large climatic gradients, such as the North American tallgrass prairie. Here, we used a short-term mesocosm experiment to evaluate how seed provenances (Local Northern region, Non-Local Northern region, Non-Local Southern region) of 10 native tallgrass prairie plants (four forbs, two legumes, and four grasses) responded to warmer conditions predicted in the future and how soil microbial communities from those three regions influenced these responses. Warming and seed provenance affected plant community composition, and warming decreased plant diversity for all three seed provenances. Plant species varied in their individual responses to warming, and across species, we detected no consistent differences among the three provenances in terms of biomass response to warming and few strong effects of soil provenance. Our work provides evidence that warming, in part, may reduce plant diversity and affect restored prairie composition. Because the southern provenance did not consistently outperform others under warming and we found little support for the “local is best” paradigm currently dominating restoration practice, identifying appropriate seed provenances to promote restoration success both now and in future warmer environments may be challenging. Due to the idiosyncratic responses across species, we recommend that land managers compare seeds from different regions for each species to determine which seed provenance performs best under warming and in restoration for tallgrass prairies.