The objectives of this paper were to study the association between the prevalence of uncontrolled hypertension (PUHT) and stroke mortality at the ecological level, in nine geographical areas of Catalonia (Spain); to develop an ecological regression model and to assess its ability to predict crude stroke mortality rate (CMR) from the PUHT. The regression equation obtained for the population older than 25 yrs was CSMR x 10(3) = 0.67035 + 4.94752 PUTH x 10(-2). The ecological risk ratio was 8.38 and the ecological attributable proportion 71.1%. The CSMR estimation obtained by applying the model in a concrete case was close to that observed. The results support an ecological association between the studied variables. The ecological model can be useful in the assessment of observed changes in health problems and risk factor levels in the community. It could also be used in the evaluation of intervention programmes.