This meta-analytic study investigates the effectiveness of different interventions in alleviating loneliness among Chinese older adults aged 50 years and above. We searched eight English databases, four Chinese databases, and grey literature. Thirty-four studies, including four randomized controlled trials (RCTs) and 30 quasi-experimental studies, were eventually included in the meta-analysis (n = 3843). Quality appraisal indicated risks of bias in the included studies. The pooled effect size was large and significant (Hedge's g = 0.84, 95% CI [0.54, 1.15]), indicating the effectiveness of interventions in reducing loneliness. However, the effect size may be overestimated due to publication bias. Moderation analyses showed significant differences in effect sizes by study designs and regions of studies. Network meta-analysis (NMA) indicated that hybrid and psychological interventions appeared to be advantageous over others. In addition, group-based delivery modes can add extra benefits to the interventions. This study adds to the knowledge of the effectiveness of current interventions in reducing Chinese older people's loneliness. However, the findings need to be interpreted with caution due to the relatively low study quality, considerable heterogeneity, and publication bias. Despite the limitations, this study offers valuable insights for future research, practice, and policy-making in reducing older people's loneliness.