Based on the meteorological data and frost damage observation data in Shangqiu during 1956-2015, meteorological conditions for the occurrence of late frosts were analyzed firstly, and then partial correlation and linear regression model of grass surface temperature (GT) and air temperature (AT), ground temperature (ST), relative humidity (RH), wind speed (WS) were studied to explore the relationship between meteorological factors and late frost damage. The results showed that the accumulated temperature (CT) in winter and spring increased year by year at 4.35℃·d·y~(-1) with the climate warming. Late frost injury to winter wheat was closely related to the antecedent precipitation and accumulated temperature in addition to the cold air activity. If previous accumulated temperature was too high or the amount of precipitation was less, frost damage was prone to occur. With the development of young ear, its sensitivity to sub-freezing temperature increased, and the minimum temperature causing frost injury had a tendency to rise. Air temperature, ground temperature, average relative humidity and average wind speed were significantly correlated with grass surface temperature (P<0.01), with an order of ground temperature>air temperature>average relative humidity>average wind speed. It could be seen that the average relative humidity and average wind speed also played important roles to the occurrence and severity of frost damage, in addition to air temperature and ground temperature. The multiple linear regression model of the four meteorological factors and grass surface temperature was established by following formula: Y=0.558ST+0.482AT+0.087RH+1.304WS-12.704. The model has been tested by 0.01 significant levels and could predict grass surface temperature very well. Furthermore, reliable frost damage assessment results could provide for areas where the grass surface temperature sensor was not installed.