The Qinghai-Tibet Plateau (QTP), a vital ecological security barrier in China, faces significant degradation due to climate change and anthropogenic activities. However, except grassland ecosystem, only a few studies have focused on a large forest ecosystem in China. This study detects forest degradation using some important landscape metrics (large patch index, aggregation index, Shannon’s evenness index, Shannon’s diversity index), maximum entropy model, and geographical detector. Results show that potential forest area on QTP is 29.30×10 ⁴ km ², with an improvement potential of 18% in the south, while 5.53×10 ⁴ km ² forest is scattered in the northeast QTP, which is fragmenting and degrading. Existing landscape metrics reveal a current large patch index (LPI) of 18.11, indicating some areas of forest ecosystem have a certain scale of the continuous habitat, but most areas may be composed of small patches or highly scattered habitat, and aggregation index (AI) of 52.54 indicating the distribution of vegetation landscape in this area is relatively balanced, with no obvious aggregation or discrete phenomenon, and predicted LPI and AI were increased to 25.95 and 61.92 with the potential forest ecosystem, respectively. Moreover, the geographic detector model, which was used to reveal the driving forces of forest change, showed among several environmental and human factors, biomass contributed most to the potential forest, with a permutation importance of 45.9%. Human activities especially road construction are the main drivers of degradation, while natural factors such as sand and clay also play a role in degradation. To address this problem, intensified afforestation is recommended for the southern area. Afforestation will enhance vegetation coverage and biomass and restore the forest ecosystem, strengthening the Plateau’s ecological barrier functions. Results presented here may provide a scientific basis for restoring QTP’s forest ecosystem and help mitigate its degradation.