The forest-steppe transition ecozone in the Daxing'an Mountains of Northeast China is highly sensitive to climate change and human activities, and it's an important ecologically fragile area of China. Studying the spatial-temporal changing and driving factors of vegetation index, which characterizes the ecological environment changing is the theoretical basis of policymaking and improving ecological environment. Therefore, we used regional analysis and empirical analysis to analyze the Normalized Difference Vegetation Index (NDVI) dynamic and influence factors, and used variable importance in projection index to quantitatively analyze the relative importance of environmental factors and human activity factors. The results showed that, the NDVI showed a degradation trend (-0.02/10a) during 1982-2015, the minimum temperature, the mean temperature and the maximum temperature showed a warming trend of 0.13℃/10a, 0.16℃/10a and 0.20℃/10a, respectively, while the annual average precipitation showed a downward trend with time (-16.3 mm/10a). For the correlation between climate factors and the NDVI, the correlation between NDVI and the minimum temperature, the mean temperature and the maximum temperature was 0.01, 0.02 and 0.04, respectively. The impact of potential evapotranspiration on NDVI varies across regions, when precipitation within 200-400 mm potential evapotranspiration significantly negatively correlated with NDVI (the proportion was 18.60% of the study area), while precipitation above 400 mm potential evapotranspiration significantly positively correlated with NDVI for 16.01% of the study area. Precipitation positively correlated with the NDVI (R=0.03), including 19.55% of significant positive correlation and 5.31% of significant negative correlation pixels. Precipitation was the dominant influencing factor of the NDVI in Chen Barag Banner, Xin Barag leaf Banner and Ewenki Autonomous Banner (where grass is the main vegetation). Cloud coverage was the dominant influencing factor of eastern woodland. In addition, the empirical analysis results showed that human activity factors (total cultivated area and livestock density) had higher explain rate than temperature and precipitation, and the average importance index of human activity (VIP _{human activity factor}=2.48) was higher than that of climate factor (VIP _{climate factor}=0.80), besides the precipitation explaining rate and VIP value was higher than of temperature. Therefore, under the background of climate warming, the climate of forest-steppe transition ecozone in the Daxing'an Mountains is showing a trend of warming and aridity. The influence of human activities on vegetation dynamic cannot be ignored. Rationally regulated agriculture and livestock is an important way to improve the stability and sustainability of vegetation ecosystem in the forest-steppe ecotone.