Pea (Pisum sativum L.) is an important annual legume crop grown in temperate regions for its high seed nitrogen concentration and environmental benefits. In the recent climate warming, a subtle evolution of the winter crop frost risk was observed: a paradoxical increase of frost stress events and a frost stress intensity decrease (Castel et al. 2017). Such results are questioning the future winter frost risk for peas. We assessed the winter frost damage evolution along 2006 to 2100 in Burgundy-Franche-Comté (a French region - western part of Europe). The approach is based on the combination of i) a dynamical downscaled climate data of two RCP trajectories (4.5 and 8.5) (Boulard et al. 2016) and ii) a winter frost stress model calibrated and validated for pea (using varieties with different frost resistance levels and acclimation rates) (Lecomte et al. 2003; Castel et al. 2017). Our results show that frost risk will not disappear with warming climate (Fig. 1). Compared to the historical period (1980-2005), the frost risk for the pea variety with a frost resistance level of -13°C will increase along the near future period (2020-2050) for RCP 8.5: with an increase of both the median and the spread of the cumulative frost degree days (Fig. 1B). With a highest warming along the far future period (2070-2100) for RCP 8.5, the results show a significant decrease of the cumulative frost degree days compared to the near future and the historical periods, but the frost risk will persist (Fig. 1B). It suggests that frost risk will significantly increase for an extended winter warming below + 2°C, while it will decrease when this threshold will be overpassed (Fig. 1). The figure 2 depicts the evolution of the two components of the frost stress with warming: intensity and number of the frost stress events. The increase of the cumulative frost degree days in the near future period (2020-2050) for RCP 8.5 is determined by the increase of frost stress events intensity (Fig. 2A). By contrast the number of frost stress events slightly decrease during this period (Fig. 2B). This result differs from the past evolution of these components with the observed warming from 1961 to 2018 (Castel et al. 2019) and suggests a change in the winter frost risk structure. For the end of the century (period 2070-2100) and for the RCP 8.5, both intensity and number of the frost stress events will decrease (Fig. 2). Finally the projections show a contrasted geography of the frost risk evolution. This geographic trend depends on the frost resistance level and acclimation rate of the pea variety. Our results seem to confirm subtle evolutions of winter climate warming dynamics revealed by the change in the pea crop frost risk structure. Moreover, this work provides leads for breeding and crop management techniques strategies for winter pea adaptation to climate change to avoid the detrimental effects of frost while taking advantage of the potential of this crop.