Background Top-down implementation strategies led by researchers often generate limited or tokenistic community engagement. Co-creation, a community engagement methodology, aims to create a shared leadership role of program beneficiaries in the development and implementation of programs, and encourages early and deep involvement of community members. We describe our experience using a four-stage co-creation approach to adapt and implement a sexually transmitted diseases (STD) testing intervention among men who have sex with men (MSM) in China. Methods We adapted a four-stage approach to co-creation. First, we conducted a needs assessment based on our prior work and discussions with community members. Second, we planned for co-creation by establishing co-creator roles and recruiting co-creators using both stratified convenience and opportunistic sampling. Third, we conducted co-creation via hybrid online/in-person focus groups (four multistakeholder groups and four MSM-only groups). Finally, we evaluated validity of the co-creation process through qualitative observations by research staff, analyzed using rapid qualitative analysis, and evaluated co-creator experience through post-discussion survey Likert scales and open-ended feedback. Results Needs assessment identified the needs to adapt our STD intervention to be independently run at community-based and public clinics, and to develop explanations and principles of co-creation for our potential co-creators. In total, there were 17 co-creation members: one co-creation lead (researcher), two co-chairs (one gay influencer and one research assistant), eight MSM community members, four health workers (two health professionals and two lay health workers) and two research implementers and observers. Co-created contents for the trial included strategies to decrease stigma and tailor interventions to MSM at public STD clinics, strategies to integrate STD testing services into existing community-led clinics, and intervention components to enhance acceptability and community engagement. Our evaluation of validity identified three main themes: challenges with representation, inclusivity versus power dynamics and importance of leadership. Surveys and free responses suggested that the majority of co-creators had a positive experience and desired more ownership. Conclusion We successfully adapted a structured co-creation approach to adapt and implement an STD testing intervention for a vulnerable population. This approach may be useful for implementation, and further research is needed in other contexts and populations.