In recent years the United Nations initiative on Reducing Emissions from Deforestation and forest Degradation (REDD+) program gained increasing attention in the policy arena, representing a valuable incentive for developing countries to take actions to reduce greenhouse gas emissions and at the same time promote sustainable forest management and improve local livelihoods. To design an effective REDD+ implementation plan at the local level it is crucial to make an in-depth analysis of the international and national requirements, analyse the forest change processes and related drivers at sub-national scale, and assess the local management options and constrains to ultimately select the appropriate policy mix and land management interventions. The present chapter first describes the state and historical changes of forests in Vietnam, identifies the direct and underlying drivers of deforestation and forest degradation at national scale, discusses the role of forests for climate change mitigation and indicates the key activities for reducing carbon emissions in Vietnam. Second, the main biophysical parameters and processes are assessed at sub-national scale for the Vu Gia Thu Bon river basin. The land cover and carbon stocks are mapped and quantified for the year 2010 and the land cover change and related carbon emissions are estimated for the period 2001–2010, allowing to model the land cover change and predict deforestation risks until the year 2020. Among the areas at higher risk of deforestation, the Tra Bui commune located in Quang Nam province is selected to design a sub-national REDD+ implementation plan in the third part of the chapter. The plan is based upon an in-depth analysis of the local context and land cover change dynamics, the local drivers of deforestation and a conducted stakeholder involvement process in the commune. Based upon this analysis, the last section provides recommendations about the local land management strategies that could be introduced in the commune and discusses the policy interventions are likely to enable their implementation.