BACKGROUND: Traditional partner notification methods have been implemented for HIV-infected patients, as well as HIV treatment, in order to identify people at risk of HIV infection, especially men who have sex with men (MSM), since they are more likely to have casual sex partners. These traditional methods have some limitations. OBJECTIVE: Our study focused on developing an mHealth app to improve partner notification in practice for MSM; the study then focused on evaluating the effects of the app. METHODS: We developed an mHealth app with different modules using Java and HTML5 and tested it in an MSM community to prevent HIV transmission. The HIV incidence stratified by different follow-up periods were calculated. Poisson regression and social networks were used to estimate the risk ratios and to identify the connection among MSM, respectively. RESULTS: In addition to the partner notification module, which is the kernel of the app, we developed a test result self-query module to enable MSM to get their approved test results in a timely manner, a prompt and warning module to alert users to protect themselves from high-risk conditions, and a health education module to teach users more skills regarding HIV/AIDS prevention. Over a 1-year duration, a total of 3186 MSM used the app, of which 678 had at least two HIV test results since becoming app users; they were included in the final analysis. Among 678 users, a total of 6473 self-queries and 623 partner notifications were recorded, which identified 180 social networks of MSM app users. Those who used the partner notification function were more likely to have self-queries (P<.001). The 678 MSM app users covered 296.47 person-years and contributed to 20 HIV seroconversions; the cumulative HIV infection incidence was estimated as 6.75 per 100 person-years (95% CI 4.38-10.01). We found that the longer the app was used, the lower the HIV incidence (>5 months vs ≤5 months: 2.22 per 100 person-years vs 6.99 per 100 person-years; risk ratio 0.32, 95% CI 0.12- 0.87). CONCLUSIONS: The app developed in this study is consistent with the World Health Organization's sensitivity and confidentiality recommendations; it has the potential to reduce the risk of HIV infection among MSM.