Current randomized trial evidence of the effects of physical activity interventions in older adults is mainly from developed countries, with little reliable evidence from low- and middle-income countries, such as China, where race, culture, and lifestyles differ substantially from those in Western populations. To evaluate the effects of a multilevel intervention for increasing leisure-time activity levels in Chinese older adults. This cluster randomized trial was conducted from May 2021 to May 2023, including an 8-week intervention period and a follow-up period of 24 months. Eight villages in China were randomly assigned to the intervention (4 villages) or the control (4 villages) group. Potentially eligible participants were 60 years or older. A total of 511 older adults from the selected 8 villages were enrolled. The intervention group received an 8-week socioecological model-based intervention, comprising individual, interpersonal, and community-level components, whereas the control group did not. The primary outcome was the change in leisure-time activity at 8 weeks, measured by the Physical Activity Scale for the Elderly (PASE). Possible PASE scores for leisure-time activity range from 0 to 502, with higher scores indicating higher activity levels. A total of 511 older adults from 8 villages were recruited and randomly allocated to the intervention (240 participants, 4 villages) or control (271 participants, 4 villages) groups. The mean (SD) age was 70.95 (5.72) years, and 284 (55.6%) were female participants. Seven serious adverse events (unrelated deaths) were reported. Participants in the intervention group increased leisure-time activity more than those in the control group, with a mean difference in PASE scores of 13.74 points (95% CI, 8.58-18.91 points) between the groups at 8 weeks (P < .001). Significant differences in leisure-time activity were also found over 24 months (mean difference in scores at 4 weeks, 11.66 points; 95% CI, 6.41-16.90 points; P < .001; at 6 months, 12.35 points; 95% CI, 7.19-17.50 points; P < .001; at 12 months, 11.55 points; 95% CI, 6.32-16.78 points; P < .001; and at 24 months, 14.51 points; 95% CI, 9.28-19.75 points; P < .001). In this cluster randomized trial, the multilevel intervention was effective in promoting leisure-time activity over 24 months of follow-up in older adults from China. This finding suggests that implementation of such interventions could be an important step in addressing physical inactivity in older adults in low- and middle-income countries. Chinese Clinical Trial Registry Identifier: ChiCTR2100045653.