Freezing rain has been normally considered to be composed of supercooled raindrops when the 2 m air temperature (hereafter Tas ${T}_{as}$) is below freezing. However, according to a statistical survey of freezing rain observations in China from 2000 to 2019, we find that there were 656 events that occurred at Tas ${T}_{as}$ greater than 0°C (hereafter warm freezing rain and denoted by WFR), which account for 7% of the total freezing rain observations. Additionally, nearly 3% (266 observations) of freezing rain events occurred when the near‐surface wet‐bulb temperature was greater than 0°C. The modeling and sensitivity experiments on the nonequilibrium raindrop temperature (hereafter Tr ${T}_{r}$) show that the temperature difference between raindrops and the atmosphere is the main cause of WFR. The magnitudes of the normalΔTar ${\Delta}{T}_{ar}$ (difference between raindrop temperature Tr ${T}_{r}$ and air temperature Ta ${T}_{a}$) and normalΔTwr ${\Delta}{T}_{wr}$ (difference between Tr ${T}_{r}$ and wet‐bulb temperature Tw ${T}_{w}$) are determined by the raindrop diameter D, temperature lapse rate Г, and relative humidity RH. Increases of D and Г, and a decrease of RH enhance normalΔTar ${\Delta}{T}_{ar}$ and normalΔTwr ${\Delta}{T}_{wr}$ and thus the occurrence of WFR. Further simulations of 4 idealized and 370 real sounding profiles reveal that either the Ta ${T}_{a}$ or the Tw ${T}_{w}$ cannot properly distinguish the WFR events. When considering the temperature difference between raindrops and the atmosphere, the WFR can form by the “melting of solid hydrometeors” or “supercooled warm rain process.” This study can also deepen our understanding of the conditions of WFR and freezing rain formation at different altitudes.