Vegetation phenology, considered as an ideal indicator of climate change, is a key aspect of plant life strategy that determines the ability to capture seasonally variable resources. As an important climatic factor, temperature substantially impacts vegetation phenology in many parts of the world and shows dominant effects in the plateau area particularly. During the past few decades, warming in Tibetan Plateau is believed to have been higher than that for the rest of the globe, particularly in the trend that daily minimum temperature (T_min) increased significantly faster than daily maximum temperature (T_max). However, the alpine vegetation phenological response to asymmetric increases in T_min and T_max is unknown. The underlying mechanism provides the novel insight that asymmetric changes in T_min and T_max should be explicitly considered in dynamic global vegetation models to elucidate the effects of climate change on vegetation phenology. Owing to the lack of in-situ observations, assessments of vegetation phenological changes in the Tibetan Plateau are mainly dependent on remote sensing data. However, because of relatively coarse spatial and temporal resolutions, as well as many other factors (data sources, retrieval methods), remote sensing information introduces great uncertainty to vegetation phenology determinations. Great controversy has arisen about the satellite-based delayed trend of spring phenology induced by current global warming in the Tibetan Plateau, as there is an absence of adequate field observations. Hence, based on in-situ observations of the dominant species Kobresia humilis in the Qinghai Lake region, the largest inlet lake in the northeastern part of the Tibetan Plateau, we use partial least squares (PLS) regression to reveal the characteristics of plants phenology and their variations, as well as the potential response mechanisms of alpine plants to asymmetric increases in T_min and T_max on a daily scale. The results indicated that (1) from 1997 to 2010, the mean temperature increased in the Qinghai Lake area at a rate of 0.5℃/10 a, the mean minimum temperature also significantly increased at a rate of 0.7℃/10 a (P < 0.05), and the mean maximum temperature showed no obvious change. (2) From 1997 to 2010, the average green-up and leaf coloring dates of Kobresia humilis were April 18th and October 2^nd, respectively. With delayed green-up date and early coloring date, the length of the growing season decreased. (3) The critical periods of temperature for the green-up dates of Kobresia humilis were January and March-April. To be specific, warming in January postponed the fulfillment of chilling requirement during the dormant period and consequently delayed the vegetation green-up date. In contrast, temperature increases in March-April accelerated the accumulation of heat and thus resulted in an early onset of spring green-up. In terms of the end of growing season, the critical periods of temperature for coloring dates were early-to mid-July and August. Specifically, warming in these periods shifted the vegetation coloring date to occur earlier. (4) Based on the results of PLS regression and correlation analysis, T_min significantly impacted plants phenology during each critical period; however, T_max only showed a significant impact on the coloring dates in August. Conclusively, T_min played a dominant role in controlling the phenology of Kobresia humilis in the Tibetan Plateau.