Adaptive management of natural resources explicitly incorporates learning into a process of science-based decision making. First proposed in the late 1970s, over time, adaptive management has become an increasingly common approach to management of complex social-ecological systems characterized by high uncertainty and conflict. Rangelands used for livestock production are one such system. Management outcomes of any given livestock operation are a unique combination of the local ecology found in a particular place, the management goals and approach of the operator and regulators, and regulatory requirements. Scholars have hypothesized that adaptive management can reduce conflict by allowing experimentation and increased management flexibility when there are competing viewpoints. We used a change to adaptive management for grazing allotment administration implemented by the US Forest Service in Arizona and New Mexico to evaluate this research question. We interviewed ranchers and Forest Service employees in Arizona and New Mexico to understand how they define adaptive management, what changes they experienced due to the shift to adaptive management, and their perceptions of the impacts of adaptive management. A decade after its implementation, our data show that adaptive management alone does not reduce conflict but can help reduce conflict where trust between permittees and the US Forest Service is strong. Where trust is lacking, adaptive management may strengthen existing patterns of conflict. These findings have important management implications for the use of adaptive management in the administration of public lands and as a conflict management tool.