Developing flowering phenological models is conducive to the accurate simulation of flowering periods of ornamental plants and could provide basis for seasonal flowering tourism events. We investigated four widespread ornamental plants with high aesthetic value, namely, Amygdalus persica, Armeniaca vulgaris, Cercis chinensis and Syringa oblate. Using the first flowering date (FFD) and end of flowering date (EFD) data of these species and corresponding meteorological data at 42 sites, we developed and validated the spatiotemporal model of FFD and EFD and reconstructed data series of FFD, EFD and flowering duration (FD)of the four species over their distribution area from 1962 to 2013. At last, we analyzed spatiotemporal patterns of mean phenophases and phenological changes. The results showed that the spatiotemporal model was able to simulate the flowering phenology accurately on large spatial and temporal scales with root-mean- square-error of about 4 - 6 days. The simulated mean FFD, EFD and FD followed certain geographical gradients. Latitude was the major factor influencing flowering phenology. The FFD and EFD were delayed by 1.23 - 4.46 days and the FD was extended by 0.07-1.47 days per degree increase of latitude. Over the past 50 years, the mean FFD and EFD of all species became earlier with a rate of 0.95 - 1.61 days decade The advance of S. oblate FFD and EFD exhibited no obvious spatial pattern, while the changes of FFD and EFD for the other three species showed a noticeable spatial variation with clearer advance in the north than in the south. The FD of S. oblate extended by 0.20 days decade~(-1),while the other three species showed very weak trends of-0.01 - 0.07 days decade~(-1). The changes of FD showed strong spatial heterogeneity. These results provided a scientific basis for simulating flowering phenology of typical ornamental plants and assessing their phenological responses to climate change.