We examined changes in the abundance of immature Aedes aegypti at the household and water storage container level during the dry-season (June-July, 2008) in Tri Nguyen village, central Vietnam. We conducted quantitative immature mosquito surveys of 171 containers in the same 41 households, with replacement of samples, every two days during a 29-day period. We developed multi-level mixed effects regression models to investigate container and household variability in pupal abundance. The percentage of houses that were positive for I/II instars, III/IV instars and pupae during any one survey ranged from 19.5-43.9%, 48.8-75.6% and 17.1-53.7%, respectively. The mean numbers of Ae. aegypti pupae per house ranged between 1.9-12.6 over the study period. Estimates of absolute pupal abundance were highly variable over the 29-day period despite relatively stable weather conditions. Most variability in pupal abundance occurred at the container rather than the household level. A key determinant of Ae. aegypti production was the frequent filling of the containers with water, which caused asynchronous hatching of Ae. aegypti eggs and development of cohorts of immatures. We calculated the probability of the water volume of a large container (>500 L) increasing or decreasing by ≥20% to be 0.05 and 0.07 per day, respectively, and for small containers (<500 L) to be 0.11 and 0.13 per day, respectively. These human water-management behaviors are important determinants of Ae. aegypti production during the dry season. This has implications for choosing a suitable Wolbachia strain for release as it appears that prolonged egg desiccation does not occur in this village.