A four-component, non randomized, primary-care based intervention for malaria control was implemented in Buenaventura, Colombia.To evaluate the effect of the applied intervention on knowledge about malaria, attention of febrile events and frequency of malaria occurrence in three communities of Buenaventura.A post-intervention evaluation with a non-equivalent control group was performed. Two non-intervened groups (those residing more and less than six months, respectively, in the area) and one intervened group were identified. We interviewed 661 women household heads. Contact was defined as having been exposed to at least one of the four intervention components.Fourteen percent of the respondents had contact with the intervention. The attention of a febrile episode was better in those who had contact with the intervention than in the nonintervened ones who had resided in the area for more than six months. Those without contact and with less than six months stay in the area reported lower use of bed-nets (OR:0.46; 95% CI:0.23-0.93) and less fumigation practices (OR:0.38; 95% CI:0.19-0.75). The analysis of the malaria case trend showed a reduction in the proportion of cases contributed by the intervened communities, from 25% to 17%, six years after the intervention.An educational strategy is effective to enhance knowledge and modify the practices of the urban population of Buenaventura with respect to malaria.