Simulating temporal and spatial relationships between occurrence dates of plant phenophases and climatic factors is crucial not only for predicting phenological responses to climate change but also for identifying the carbon-uptake period and examining the seasonal exchanges of water and energy between land surface and atmosphere. The latter in turn affects the global carbon cycle and climate change. In order to reveal the spatial pattern and its ecological mechanism of temporal variation of plant phenology and plant phenology response to climate change in China's temperate zone, we used Salix matsudana's phenology data of first leaf unfolding (LU), first flowering (FF), fruit maturing (FM), first leaf coloration (LC) and the end of leaf fall (LF) at 52 stations during 1986-2005 to analyze the linear trend of phenological time series and identify the response of phenological occurrence dates to interannual temperature variations by establishing daily mean air temperature-based temporal phenology models. During the research period, regional mean occurrence dates of Salix matsudana's LU, FF and FM significantly advanced at average rates of -4.2, -3.8 and -3.3 days per decade, respectively, whereas regional mean occurrence dates of Salix matsudana's LC and LF indicated a nonsignificant delayed trend and a significant delayed trend at an average rate of 2.4 days per decade, respectively. At single stations, occurrence dates of Salix matsudana's LU, FF and FM significantly advanced at 40%, 41% and 29% of stations, respectively; occurrence dates of Salix matsudana's LC significantly advanced at 17% of stations and delayed at 19% of stations; occurrence dates of Salix matsudana's LF significantly delayed at 23% of stations. Spatial series of linear trends in occurrence dates of Salix matsudana's LU, FF and FM at all stations correlate negatively with spatial series of linear trends in corresponding daily mean air temperatures during the optimum length periods, namely, the quicker the preceding air temperatures increased at a station, the quicker the phenological occurrence dates advanced at the station. With respect to phenological response to interannual temperature variations, a 1℃ increase in regional mean spring air temperatures during the optimum length periods may induce an advancement of 3.08 days, 2.83 days and 3.54 days in regional mean occurrence dates of Salix matsudana's LU, FF and FM, respectively, whereas a 1℃ increase in regional mean autumn air temperatures during the optimum length periods may cause a delay of 1.69 days and 2.28 days in regional mean occurrence dates of Salix matsudana's LC and LF, respectively. At single stations, the response of occurrence dates of Salix matsudana's LU and LF to interannual temperature variations was more sensitive at warmer locations than at colder locations. Overall, simulation precision of daily mean air temperature-based temporal phenology models for spring and summer phenophases was obviously higher than that for autumn phenophases. Therefore, we constructed daily mean air temperature and daily accumulative precipitation-based autumn phenology models. The revised models significantly enhanced simulation precision of Salix matsudana's LC and LF. This indicates that occurrence dates of Salix matsudana's LC and LF were triggered by the combined influence of preceding air temperature and precipitation.