To study the effect of a community-based walking intervention on blood pressure among older people. The study design was a randomized controlled trial conducted in a rural area of Taiwan between October 2002 and June 2003. A total of 202 participants aged 60 years and over with mild to moderate hypertension was recruited. Participants randomized to the intervention group (n=102) received a six-month community-based walking intervention based on self-efficacy theory. A public health nurse provided both face-to-face and telephone support designed to assist participants to increase their walking. Control group participants (n=100) received usual primary health care. Primary outcome was change in systolic blood pressure and secondary outcomes were exercise self-efficacy, self-reported walking and diastolic blood pressure. At six-month follow-up the mean change in systolic blood pressure was a decrease of 15.4 mmHg and 8.4 mmHg in the intervention and control group, respectively. The difference in mean change between the two groups was -7.0 mmHg (95% CI, -11.5 to -2.5 mmHg, p=0.002). Improvement in exercise self-efficacy scores was greater among intervention group participants (mean difference 1.23, 95% CI, 0.5 to 2.0, p=0.001). Intervention group participants were more likely to report walking more (p<0.0005) but no differences were observed in diastolic blood pressure (p=0.19). Among hypertensive older people, a six-month community-based walking intervention was effective in increasing their exercise self-efficacy and reducing systolic blood pressure.