Restoring degraded landscapes is critical for achieving global environmental and development goals, and agroforestry is increasingly promoted as a nature-based solution to land degradation. Farmer-managed natural regeneration (FMNR) is an agroforestry-based approach for restoring degraded agricultural land and it has been widely implemented in African drylands. However, a recent systematic review found significant gaps in the evidence base for FMNR, including that its upscaling has been based on inadequate understandings of local contexts. Furthermore, studies reporting on farmer adoption of FMNR have mainly relied on quantitative data from household surveys, resulting in limited understandings of what motivates farmers who practice FMNR. This paper draws on the results of a qualitative study in northeastern Ghana to address two questions: 1) How and why do farmers practice FMNR? And 2) How does context influence farmers’ rationales for practicing FMNR? We found that farmers grounded their perspectives on the utility of FMNR in nuanced understandings of the local farming and land and tree tenure systems. The results of our study also demonstrate how farmers’ decision-making was situated within socially and agroecologically differentiated contexts, which were conditioned by long-term, multi-faceted change in the region. We conclude that in spite of the rush to scale up FMNR, more attention should be directed to assessing where, when, and for whom FMNR might be appropriate. Such assessments should be grounded in resource managers’ preferences, local agricultural and land and tree tenure systems, and the requisite biophysical conditions for FMNR. To support these efforts, we propose an FMNR suitability assessment framework, based on our findings and those from related studies. As landscape restoration is scaled up globally, initiatives should be informed by evidence demonstrating how and why resource managers might practice a restoration activity as well as how context influences their choices.