Conservation of natural habitats in human-dominated landscapes is critical for halting biodiversity loss. Maintaining habitat quantity and connectivity requires landscape-level collective action, which results from environmental decisions made by individual land owners. We investigate how individual decision making in a rural collective translates into quantitative differences in landscape-level environmental outcomes. Behavioral science has become a critical domain of knowledge in conservation, but little attention has been paid to how multiple behavioral drivers determine the success of collective environmental action. We developed a social-ecological model for landscape-level conservation using a detailed data set of 600 land owners in New Zealand. With the model, we tested whether the effect of social influence networks on collective conservation action was altered by their interplay with land owners’ personal characteristics, connections to cross-scale actors and local environmental contexts. Interactions between multiple behavioral drivers determined the environmental outcomes of collective action in unexpected ways by modifying, muting or amplifying the effects of single drivers. Importantly, we detected a social-ecological mechanism for rapid change in the extent of protected habitats, which can explain highly successful or failed environmental outcomes of collective conservation. Further, when environmentally desirable and undesirable behaviors spread simultaneously through the social network, homophily and network cohesion hinder desirable environmental outcomes. This effect can be modified by other drivers such as social responses to local environmental change. Thus, understanding how the antagonistic and synergistic effects of behavioral drivers can be best utilized in conservation will benefit biodiversity and ensure benefits that humans obtain from biodiversity.