This study analyzed the relationship between blood pressure and all-cause mortality according to objectively measured walking speed in a Mediterranean population-based sample of older persons. We used data from the longitudinal 'Peñagrande' Cohort Study, initiated in 2008 in a sex- and age-stratified random sample of 1250 people aged ⩾65 years living in Madrid (Spain). A total of 814 individuals participated in the first study wave. The average of two standardized blood pressure readings was used. Walking speed was measured over a 3-m walk and classified as faster (⩾0.8 m s(-1)) or slower. A total of 314 individuals were slower walkers, 475 were faster walkers and 25 did not complete the walk test. Cox proportional hazards models stratified by walking speed were used to assess the association between blood pressure and all-cause death. Non-linear relationship between BP and mortality was explored by a restricted cubic spline analysis. There were 171 deaths from study entry through 31 March 2013. Systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mmHg were associated with higher mortality than blood pressure values above 140 and 90 mm Hg, respectively, but this association reached statistical significance only for systolic blood pressure and only in the slower walkers. In conclusion, systolic blood pressure levels <140 mm Hg were found associated with higher risk of total mortality among slower walkers in an old Spaniard population cohort.