Through analyzing the factors affecting the extreme temp eratures of 11 cit ies in Hebei Province and integrating the merits of experiential forecast and nu merical forecast, the forecasting equations of extreme temperatures of these 11 cities are established by using the successive regression method. The results sh ow that this method can score 80 in temperature objective forecasting, and even more effective when the amplitude of temperature change is larger during spring and fall, as well as in transitional and break weather conditions.