Abstract Background and Aims Spring phenological change of plants in response to global warming may affect many ecological processes and functions. Chilling temperature regulates budburst date by releasing dormancy. However, whether freezing temperatures (<0 °C) contribute to dormancy release remains of debate. Our poor understanding of the role of chilling makes estimating shifts in budburst date difficult. Methods A 2-year chilling–forcing experiment was explicitly designed to test the effects of chilling temperatures on dormancy release of nine temperate woody species in Beijing, China. A total of 1620 twigs were first exposed to a wide range of temperatures (−10 to 10 °C) with different durations and then moved to growth chambers. Based on budburst data in experimental conditions, we examined whether freezing temperatures are effective on dormancy release. We also developed a new framework for constructing chilling functions based on the curve between chilling duration and forcing requirement (FR) of budburst. The chilling function derived from this framework was not affected by experimental forcing conditions. Key Results We demonstrated that freezing temperatures down to −10 °C were effective in dormancy release. The rate of dormancy release, indicated by the rate of decay in the chilling duration–FR curve, did not differ significantly between chilling temperatures in most cases, although it exhibited a maximum value at 0 or 5 °C. The chilling function-associated phenological models could simulate budburst date from independent experimental and observational data with a mean RMSE of 7.07 d. Conclusions The effective freezing temperatures found here are contrary to the well-known assumption of <0 °C temperature generally not contributing to accumulated chilling in many previous chilling functions. A chilling function assuming that temperature below an upper temperature threshold has the same effects on dormancy release could be adopted to calculate chilling accumulation when using experiments to develop spring phenological models based on the chilling–forcing relationship.