Environmental co-governance depends not only on the local stability of collaboration but also on whether ENGO-based collaborative participation is attainable from a broad range of initial conditions. This study develops a tripartite evolutionary game model involving local governments, environmental NGOs (ENGOs), and the public, and uses basin-of-attraction analysis to examine the global attainability of high-participation environmental co-governance. The model combines replicator dynamics, Jacobian-based local stability analysis, threshold conditions, numerical simulation, and basin-of-attraction estimation. The results show that collaborative stability is conditional: high-participation co-governance emerges only when institutional support, ENGO participation incentives, and public cooperation conditions jointly exceed critical thresholds. Institutional support is more effective when it improves coordination capacity and reduces implementation friction, whereas unconditional subsidies have ambiguous effects because they increase ENGO incentives while also reducing the government’s relative payoff from strong support. Public cooperation is especially sensitive to participation burden and targeted incentives, while higher passive payoffs for ENGOs enlarge low-participation traps. The analysis is theoretical and simulation-based, informed by China’s institutionally bounded ENGO context, and is not intended as an empirically calibrated prediction for a specific locality. The findings suggest that durable environmental co-governance requires coordinated institutional arrangements that jointly strengthen governmental support, ENGO participation incentives, and public cooperation conditions.