Preventing the invasion of freshwater aquatic species is the surest way to reduce their impacts, but it is also often expensive. Hence, the most efficient prevention programs will rely on accurate predictions of sites most at risk of becoming invaded and concentrate resources at those sites. Using data from Vilas County, Wisconsin (U.S.A.), collected in the 1970s, we constructed a predictive occurrence model for rusty crayfish (Orconectes rusticus) and applied it to an independent data set of 48 Vilas County lakes to predict which of these were most likely to become invaded between 1975 and 2005. We nested this invasion model within an economic framework to determine whether targeted management, derived from our quantitative predictions of likely invasion sites, would increase the economic value of lakes in the independent data set. Although the optimum expenditure on lake protection was high, protecting lakes at this level would have produced net economic benefits of at least $6 million over the last 30 years. We did not attempt to determine the value of nonmarket benefits of protection; thus, our results are likely to underestimate the total benefits from preventing invasions. Our results demonstrate that although few data are available early in an invasion, these data may be sufficient to support targeted, effective, and economically rational management. In addition, our results show that ecological predictions are becoming sufficiently accurate that their application in management can produce net economic benefits.