BACKGROUND: The World Health Organization has recommended HIV self-testing (HIVST) as an alternative testing strategy given the limitations of facility-based testing. While the benefits of HIV self-testing have been demonstrated at the individual level among men who have sex with men (MSM), limited data exist on if this testing approach can be effectively diffused through individuals' social or sexual networks. The objectives of this study were to examine patterns and correlates of HIVST distribution within Chinese MSM's sexual networks. METHODS: Data used for this analysis was a part of the process evaluation of an HIVST intervention trial among MSM in Nanjing, China. Between May and October 2017, we enrolled 400 men into the trial. Participants assigned to the intervention group (N = 200) were given three HIVST kits at baseline and could request more during the follow-up periods. We incorporated measures for process evaluation in the self-administered online follow-up surveys. This analysis reported findings from the three-month follow-up survey in the intervention group. Frequencies and percentages were used to describe characteristics of participants who distributed kits to their sexual partners as well as patterns of distribution. Multivariable logistic regression was conducted to identify independent correlates of participants who distributed the kits. RESULTS: Of the 177 participants retained (88.5%) at the three-month follow-up, 72 (40.7%) distributed one or more kits to either primary or casual partners. About half of distributors (51.4%) gave one HIVST kit to their sexual partners while 15.3% distributed 3 or more. Over half gave these kits (58.3%) to primary sexual partners while 27.8% reported giving the kits to both primary and casual partners. About half (54.2%) of distributors used the kits together with their partners. Compared to participants who had an HIV test in the past six months, those who tested over six months ago or never tested had significantly lower odds of distributing the kits (AOR = 0.484, 95% CI: 0.250-0.983, p = 0.032). Compared to those who had not used the kits themselves, participants who did had significantly higher odds of distributing the kits (AOR = 3.345, 95% CI: 1.488-7.517, p = 0.003). Participants who reported higher HIV testing efficacy had 2.051 fold greater odds (95% CI: 1.062-3.961, p = 0.033) of distributing the kits compared to those who had lower efficacy. CONCLUSION: Our study demonstrated that a sexual network-based approach to distributing HIVST among Chinese MSM is feasible and can be a promising strategy to improve the effectiveness of HIVST programs including its reach to untested men. Such approach should be complimented by intervention components that enhance HIV testing efficacy and improve experiences of HIVST.