Abstract Background Blood pressure (BP) management in frail older people is challenging. An randomised controlled trial of largely non-frail older people found cardiovascular and mortality benefit with systolic (S) BP target <120 mmHg. However, all-cause mortality by attained BP in routine care in frail adults aged above 75 is unclear. Objectives To estimate observational associations between baseline BP and mortality/cardiovascular outcomes in a primary-care population aged above 75, stratified by frailty. Methods Prospective observational analysis using electronic health records (clinical practice research datalink, n = 415,980). We tested BP associations with cardiovascular events and mortality using competing and Cox proportional-hazards models respectively (follow-up ≤10 years), stratified by baseline electronic frailty index (eFI: fit (non-frail), mild, moderate, severe frailty), with sensitivity analyses on co-morbidity, cardiovascular risk and BP trajectory. Results Risks of cardiovascular outcomes increased with SBPs >150 mmHg. Associations with mortality varied between non-frail <85 and frail 75–84-year-olds and all above 85 years. SBPs above the 130–139-mmHg reference were associated with lower mortality risk, particularly in moderate to severe frailty or above 85 years (e.g. 75–84 years: 150–159 mmHg Hazard Ratio (HR) mortality compared to 130–139: non-frail HR = 0.94, 0.92–0.97; moderate/severe frailty HR = 0.84, 0.77–0.92). SBP <130 mmHg and Diastolic(D)BP <80 mmHg were consistently associated with excess mortality, independent of BP trajectory toward the end of life. Conclusions In representative primary-care patients aged ≥75, BP <130/80 was associated with excess mortality. Hypertension was not associated with increased mortality at ages above 85 or at ages 75–84 with moderate/severe frailty, perhaps due to complexities of co-existing morbidities. The priority given to aggressive BP reduction in frail older people requires further evaluation.