Abstract: The relationship between flowering phase and temperature was analyzed and the flowering phase of the peach tree was forecasted using the observed flowering phase in Hekou village of Dandong and daily mean temperature of the Kuandian national reference climatological station from 2005 to 2015.Three models,including active accumulated temperature model (AAT),effective accumulated temperature model (EAT) and sliding accumulated temperature model (SAT),were proposed to simulate the flowering phase based on the international accepted Spring Warming (SW) phenological model principle.The total accumulated temperature threshold before blossom and the most representative lower limit temperature of each model were determined using methods of 0.1℃ temperature interval census and mathematical deviation.Besides,the applicability of each model was tested and evaluated.The results show that SAT is the most suitable model for forecasting the peach tree flowering phase in Hekou village of Dandong with an accuracy rate of 90.91%.The accuracy of AAT and EAT are 81.82% and 63.64%,respectively.In this study,the average of maximum and minimum temperature during the period when sliding accumulated temperature passed its lower limit temperature from 2005 to 2015 was taken as an index for forecasting the peach tree flowering phase,which make the accuracy of SAT better than before.Therefore,we conclude the SAT method is a good method to forecast the flowering phase of peach tree in Dandong.