Vegetation is the most important ecosystem in the earth, which plays an important regulatory role in global climate change. Research on the temporal and spatial dynamic changes of vegetation has important scientific and practical value. Based on the normalized difference vegetation index (NDVI) of Landsat satellite remote sensing imagery extracted in 2007, 2012 and 2016 and the forest inventory data of Sanming City, the spatio-temporal changes of NDVI of Minjiang River upstream area in the last 10 years and the influencing factors were studied. The results showed that the NDVI values of the study area in 2007, 2012 and 2016 were increased gradually, and their values were 0.72, 0.75 and 0.79, respectively, which indicated that the coverage of vegetation was high, and the overall ecological quality was good. The lower NDVI values were located in the urban area, the concentration points of the township residents and near the road network. The NDVI values of different vegetation types were in a descending trend as broadleaf forest, shrub forest, Masson pine forest, bamboo forest, Chinese fir forest, miscellaneous wood forest, eucalyptus forest, other wooded land, economic forest, eucalyptus forest and non-forest land. NDVI index value was gradually increased with the increase of forest age, canopy closure and slope, while it was decreased with the increase of grade of the site quality. According to variation of changing dynamics of NDVI index with age of different types of vegetation, and the vegetation was divided into the following three types: type of steadily rising, such as broad-leaved forest and Pinus massoniana forest; type of rising fast at early stage and slow at late stage, such as eucalyptus forest; type of rising first and then fall, such as Chinese fir forest and other woodland. The NDVI value of Pinus massoniana forest was increased with the increase of slope, while the NDVI value of other forest vegetation types was risen first and then decreased with the increase of slope, and reached the highest value at the slope of 30°~40°. The NDVI index values of different vegetation types had different responses to canopy closure.