<p>Traffic fraud has become a serious threat to the live-streaming commerce ecosystem, causing adverse effects such as distorting market signals and undermining fair competition. As fraudulent tactics continue to evolve, traditional single-dimensional governance approaches face mounting challenges. This study examines the dynamic evolution of a live-streaming commerce supply chain with traffic fraud under external legal regulation and explores how internal supply chain members achieve collaborative governance within the ecosystem. To this end, we develop a tripartite evolutionary game model involving the live-streaming platform, streamers, and manufacturers, and conduct co-governance mechanism design and key-parameter sensitivity analyses. The key findings are as follows: (i) When the relative gains for streamers from traffic fraud are lower than the combined relative traffic costs and reputational value compared to non-fraudulent behavior, the system remains in a healthy and orderly state; otherwise, streamers tend to engage in traffic fraud. (ii) The platform’s and manufacturer’s reputations, regulatory (resistance) costs, compensation, technical support, penalties, and the success rate of fraud detection by each party all influence the system’s progression toward cooperative governance. (iii) As the leading actor in co-governance, the platform should collaborate with manufacturers to maintain a high reputation, which is critical for fostering cooperation. (iv) Both parties should optimize their cost structures, and the platform should accurately evaluate cost–benefit trade-offs and dynamically adjust the level of support. (v) Increasing penalties and improving fraud detection rates can motivate platforms and manufacturers to participate in governance. These findings offer actionable insights for addressing traffic fraud in the live-streaming commerce supply chain and promoting sustainable industry development.</p>