The aim of this paper is to assess the direct and indirect impacts of the agricultural extension system of Uganda, the National Agricultural Advisory Services (NAADS) program, on household agricultural income. Data from two rounds of surveys of Ugandan rural farm-households conducted in 2004 and 2007, as well as different program evaluation methods and model specifications, are used to estimate impacts and compute a rate of return. The direct and indirect impact of the program is estimated at 37–95% and 27–55% increase in per capita agricultural gross revenue between 2004 and 2007 for households participating directly and indirectly in the program, respectively, compared to nonparticipants. The rate of return on the program's expenditures is estimated at 8–49%. The program has been relatively more effective among male-headed, larger, and asset-poor households, as well as those taking up noncrop high-value enterprises and living further away from financial services, all-weather roads, and markets or located in the Eastern and Northern Regions. Policy implications of the results are drawn.