The aim was to evaluate the efficacy, safety and completion rate of 3-month, once-weekly rifapentine and isoniazid for tuberculosis (TB) prevention among Chinese silicosis patients. Male silicosis patients without human immunodeficiency virus infection, aged 18 years to 65 years, with or without latent TB infection, were randomized 1:1 to receive rifapentine/isoniazid under direct observation (3RPT/INH group) or were untreated (observation group). Active TB incidence was compared between the two groups with 37 months of follow-up. Safety profile and complete rates were evaluated. A total of 1227 adults with silicosis were screened; 513 eligible participants were enrolled and assigned to 3RPT/INH (n = 254) vs. observation (n = 259). Twenty-eight participants were diagnosed with active TB, and 9 and 19 in the 3RPT/INH group and observation groups, respectively. In the intention-to-treat analysis, the cumulative active TB rate was 3.5% (9/254) in the 3RPT/INH group and 7.3% (19/259) in the observation group (log rank p 0.055). On per protocol analysis, the cumulative active TB rates were 0.7% (1/139) and 7.3% (19/259), respectively (log rank p 0.01). Owing to an unexpected high frequency of adverse events (70.4%) and Grade 3 or 4 AEs (7.9%), the completion rate of the 3RPT/INH regimen was 54.7% (139/254). Twenty-six (10.8%) participants had flu-like systemic drug reactions; five (2.1%) experienced hepatotoxicity. Weekly rifapentine/isoniazid prophylaxis prevented active TB among Chinese people with silicosis when taken, irrespective of LTBI screening; efficacy was reduced by lack of compliance. The regimen must be used with caution because of the high rates of adverse effects. ClinicalTrials.gov number: NCT02430259