Located in a highly populated region the Kenyan Mau forest complex is increasingly exploited for cattle grazing, fuel wood collection and charcoal burning. This has big implications on the state of the forest as the understorey is disturbed and tree cover gradually diminishes. To understand the underlying forces driving forest degradation knowledge of the timing and location of anthropogenic disturbances in the forest is crucial. Using a stable forest mask for the period between 2000-2016 the analysis was limited to persistently forested areas, thus excluding deforestation. To detect forest degradation, we computed time series from the Normalized Difference Fraction Index (NDFI) which synthesizes subpixel fraction images from spectral mixture models. NDFI time series were then screened for structural changes using the data-driven Breaks For Additive Season and Trend (BFAST) model. Our results indicate an overall detection accuracy of 75.74 %. User and Producer accuracies for the mapped forest degradation class amounted 84.31% and 56.58% respectively. The nodegradation class showed user and Producer accuracies of 72.03% and 91.4% respectively. Undetected degradation patches generally remained below an area size of 0.2 ha, which reveals the importance of matching targeted degradation processes with an appropriate spatial sensor resolution. Furthermore, the year 2015 indicated an exceptional peak of degradation processes. In that year about 28 % of non-disturbed forest samples were falsely attributed as degraded. 2015 stands out with the lowest precipitation rates among all years in the study period, likely leading to the false identification of phenological anomalies as actual degradation processes. The main benefit of combining NDFI time series and BFAST is the high sensitivity towards canopy changes. Potentially, the employed change detection method can be further enhanced using additional spectral indices and breakpoint validation criteria that are more resilient to fluctuations of climate and vegetation phenology. While revealing some weaknesses of the approach the study also mapped hotspots of forest degradation in the Mau forest. These findings are relevant for the daily work of local forest services but also for future research and continued degradation mapping.