Introduction: Theoretically, it is possible to control dengue by reducing breeding sites. However in most countries, it has not been possible to obtain sustainable dengue control using only the reduction of potential breeding sites thus use of larvicides continues to be necessary. This paper presents an estimate of the rate of water replacement in residential water containers in the city of Fortaleza, northeast Brazil, and the respective duration of the effect of Diflubenzuron on Aedes aegypti immatures. Methods: Several trials were performed to estimate the duration of the effect of Diflubenzuron in simulated field conditions with replacement daily with 20% by volume of water. Eighteen containers with 20 litres of water treated with Diflubenzuron, and 6 containers with well water were used. The initial concentration of Diflubenzuron in the treated containers was 2.875 x 10-4 grams/litre. Third stages larvae were added to the containers, the first day of experiment. All the containers were inspected daily and dead larvae and pupae were removed and counted. Each trial was terminated when all larvae added to one of the groups (treated or control) had died or pupated. Results: From the 1st to the 19th day, the Diflubenzuron achieved 100% control, adjusting for the emergence of pupae in untreated containers not allowing any larvae to develop into pupae. From the 20th to the 70th day, the control ranged from 97.9% to 36.6%. Conclusions: This study reports a good larvicide effect of Diflubenzuron in the doses used by the Brazilian Programme of Dengue Control.