BACKGROUND: While the effects of initiation of antiretroviral treatment (ART) on risky sexual behavior have been extensively studied, less is known about the long-term changes in risky sexual behavior over time in resource-poor settings. METHODS: We conducted a secondary longitudinal analysis of one rural and one urban cohort of patients who initiated ART in Uganda between April 2004 and July 2007 followed up-to 2016. Data on sexual behavior were collected every 6 months for 3.5 years in individuals on ART ≥ 4 years (baseline) when a behavioral questionnaire was introduced. Risky sexual behavior was defined as sexual intercourse with ≥ 2 partners or inconsistent or no condom use in previous 6 months. We report characteristics overall, and by cohort. We used multivariable generalized estimating equations logistic regression to assess the effects of time on ART on risky sexual behavior. RESULTS: Of 1012 participants, 402 (39.8%) were urban and 610 (60.2%) were rural residents. Mean age was 42.8 years (SD 8.5). Mean duration of follow-up was 51.3 months (SD 15.3), but longer for urban than rural participants (64.5 vs 36.4 months). Risky sexual behavior declined from 33.1% at baseline to 9.6% after 3.5 years of follow-up in the rural cohort (p ≤ 0.01 for the test of trend) and was unchanged from 9.7% at baseline to 9.9% after 3.5 years in the urban cohort (p = 0.51). Receiving care at a rural clinic (aOR 4.99, 95% CI 3.64-6.84); male gender (aOR 1.66, 95% CI 1.26-2.19) and being younger (aOR 5.60, 95% CI 3.80-8.25 for 18-34 years and aOR 2.34, 95% CI 1.74-3.14 for 35-44 years) were associated with increased odds of risky sexual behavior. Not being married (aOR 0.25; 95% CI 0.19-0.34), and longer time on ART (aOR 0.71 95% CI 0.67-0.76) were associated with reduced odds of risky sex. CONCLUSIONS: We observed a decline in risky sexual behavior in rural people on long-term (≥ 4 years) ART. Rural, male and young individuals had higher odds of self-reported risky sexual behavior. ART programs should continue to emphasize risk reduction practices, especially among people receiving care in rural health facilities, males, younger individuals and those who are married.