BACKGROUND: An estimated 2·5 billion people are at risk of dengue. Incidence of dengue is especially high in resource-constrained countries, where control relies mainly on insecticides targeted at larval or adult mosquitoes. We did epidemiological and economic assessments of different vector control strategies. METHODS: We developed a dynamic model of dengue transmission that assesses the evolution of insecticide resistance and immunity in the human population, thus allowing for long-term evolutionary and immunological effects of decreased dengue transmission. We measured the dengue health burden in terms of disability-adjusted life-years (DALYs) lost. We did a cost-effectiveness analysis of 43 insecticide-based vector control strategies, including strategies targeted at adult and larval stages, at varying efficacies (high-efficacy [90% mortality], medium-efficacy [60% mortality], and low-efficacy [30% mortality]) and yearly application frequencies (one to six applications). To assess the effect of parameter uncertainty on the results, we did a probabilistic sensitivity analysis and a threshold analysis. FINDINGS: All interventions caused the emergence of insecticide resistance, which, with the loss of herd immunity, will increase the magnitude of future dengue epidemics. In our model, one or more applications of high-efficacy larval control reduced dengue burden for up to 2 years, whereas three or more applications of adult vector control reduced dengue burden for up to 4 years. The incremental cost-effectiveness ratios of the strategies for two high-efficacy adult vector control applications per year was US$615 per DALY saved and for six high-efficacy adult vector control applications per year was $1267 per DALY saved. Sensitivity analysis showed that if the cost of adult control was more than 8·2 times the cost of larval control then all strategies based on adult control became dominated. INTERPRETATION: Six high-efficacy adult vector control applications per year has a cost-effectiveness ratio that will probably meet WHO's standard for a cost-effective or very cost-effective intervention. Year-round larval control can be counterproductive, exacerbating epidemics in later years because of evolution of insecticide resistance and loss of herd immunity. We suggest the reassessment of vector control policies that are based on larval control only. FUNDING: The Fulbright Programme, CAPES (Brazilian federal agency for post-graduate education), the Miriam Burnett trust, and the Notsew Orm Sands Foundation.