In Japan, the world’s most rapidly aging country, urban farming is attracting attention as an infrastructure for health activities. In Tokyo, urban residents generally participate in two types of farming programs: allotments and experience farms. The availability of regular interaction among participants distinguishes these two programs. We quantitatively examined the difference in changes in self-reported health status between participants in these two types of urban farming. We obtained retrospective cross-sectional data from questionnaire surveys of 783 urban farming participants and 1254 nonparticipants and analyzed the data using ordinal logistic regressions. As a result, compared with nonparticipants, participants in both types of urban farming reported significantly improved self-rated health (SRH) and mental health (MH). After controlling for changes in their physical activity (PA), although participants in allotments did not report significant improvement in SRH and MH, those in experience farms did, suggesting that their health improvement was not only caused by an increase in PA but also by social interaction among participants. From the perspective of health promotion, public support is needed not only for the municipality’s allotments but also for the experience farms operated by the farmers themselves.