Introduction: Excessive sodium intake is the top dietary risk factor for mortality and non-communicable diseases such as stroke. However, whether and how a high salt diet (HSD) impacts long-term stroke outcomes - even after salt reduction/withdrawal prior to a stroke is unknown. Methods: Intracerebral hemorrhage (ICH) was modeled by blood or collagenase injections in mice. Bone marrow (BM) progenitor and immune cell compositions as well as monocyte-derived macrophage (MDM) and microglial phenotypes were profiled using flow cytometry. BM chimeras were created to test innate immune memory. Glycolysis and OXPHOS were measured by Seahorse assays. Macrophage phagocytosis was evaluated using in vivo and in vitro erythrophagocytosis assays. RNA-seq was used to analyze the progenitor transcriptome. Hematoma, neuron loss, and astrogliosis were assessed with immunofluorescence. LysM Cre/+ : NR4a1 fl/fl ; Nr4a2 fl/fl ; NR4a3 -/- mice, LysM Cre/+ : NR4a1 fl/fl mice, and celastrol treatment were used to elucidate the mechanism underlying HSD-induced trained immunity. Results: HSD impaired long-term ICH recovery despite prior salt withdrawal. HSD triggered innate immune training in BM progenitor cells by downregulating the NR4a family and OXPHOS. This training selectively compromised MDM reparative polarization without altering their initial inflammatory response in the ICH brain. Healthy mice transplanted with BM from HSD-fed mice retained training signatures, confirming a persistent form of trained immunity that originates in the BM. Lack of NR4a1 in MDMs recapitulated HSD-induced negative impacts on ICH outcomes. Notably, transplantation of healthy BM and gain of NR4a1 function enabled stroke recovery in the HSD animals. Conclusions: These findings provide the first evidence that links HSD-induced trained immunity to the acquisition of persistent dysregulated inflammatory responses and unveil NR4a1 as a potential therapeutic target.