We hypothesized that discharge SBP had different associations with outcomes in non-HFrEF (left ventricular ejection fraction ≥40%) patients with or without high blood pressure (HBP) at admission. Non-HFrEF patients hospitalized for decompensated heart failure were consecutively recruited and were categorized into HBP (admission SBP ≥130 mmHg) group and non-HBP group. The primary outcome was a composite of cardiovascular death and heart transplantation. Multivariate Cox and penalized spline analyses were used to assess the relationships between discharge SBP and outcomes. Nine hundred and sixty-four non-HFrEF patients were enrolled with a median follow-up of 71.8 months. Three hundred and sixty-five (37.9%) patients had HBP. In multivariate Cox analyses, non-HBP patients with higher discharge SBP were associated with a better outcome (per 10 mmHg increased, hazard ratio = 0.788, P = 0.001). However, an opposite relationship between discharge SBP and the primary outcome was observed in HBP group (per 10 mmHg increased, hazard ratio = 1.312, P = 0.002). Results of penalized spline regression models showed that there was a U-shaped association between discharge SBP and outcomes in the total cohort. Compared with 120 mmHg, the risk of the primary outcome increased when discharge SBP was below 99 mmHg in non-HBP group; in HBP group, a worse outcome was observed when discharged SBP was above 145 mmHg. Non-HFrEF had a U-shaped association between discharge SBP and adverse events. Such an association was modified by admission HBP. Higher discharge SBP correlated with a worse outcome in non-HFrEF patients with admission HBP, as opposed to patients admitted without HBP.