Using NDVI3g vegetation index, we defined 18 phenological metrics to investigate phenology change in the Tibetan Plateau (TP). Considering heterogeneity of vegetation phenology, we divided TP into 8 vegetation clusters according to 1:1000000 vegetation cluster map. Using partial least regression (PLS) method, we investigated impacts of climate variables such as temperature, precipitation and solar radiation on vegetation phenology. Results indicated that: (1) Turning points of the date of the start of growing season (SOS) metrics are mainly observed during 1997-2000, before which SOS advanced 2-3 d/a. Turning points of the date of the end of growing season (EOS) and length of growing season (LOS) metrics are found during 2005 and 2004-2007, respectively. Before the turning point, EOS has a delayed tendency of 1-2 d/10a, and LOS has a lengthening tendency of 1-2 d/10a. After the turning point, the tendency of SOS and EOS metrics is questionable. Meanwhile, lengthening of LOS is not statistically significant; (2) Alpine meadows and alpine shrub meadows are subject to the most remarkable changes. Lengthening LOS of alpine meadow is mainly due to advanced SOS and delayed EOS. Nevertheless, lengthening LOS of alpine shrub meadow is attributed mainly to advanced SOS; (3) Using PLS method, we quantified impacts of meteorological variables such as temperature, precipitation and solar radiation on phenology changes of alpine meadows and alpine shrub meadows, indicating that temperature is the dominant meteorological factor affecting vegetation phenology. In these two regions, autumn of last year and early winter temperature of last year have a positive effect on SOS. Firstly, increased temperature in this period would postpone last year's EOS, and hence indirectly delay SOS of the current year; Secondly, warming autumn and early winter have the potential to negatively impact fulfilment of chilling requirements, leading to delay of SOS. Except summer, minimum temperature has a similar effect on vegetation phenology, when compared to average and maximum temperature. Furthermore, precipitation effects on phenology fluctuate widely across different months. Precipitation of the autumn and winter/spring of the last year has a negative/positive effect on SOS. Besides, precipitation acts as the key driver constraining vegetation growth in August, during which precipitation has a positive impact on EOS. Therefore, solar radiation can exert impacts on vegetation phenology mainly during summer and early fall. Our research will provide a scientific support for the improvement of vegetation phenology model.