Wildfire is a necessary disturbance for forest ecosystem's regeneration and succession, but frequent and large fires threaten human life and activities. Forecasting forest fires is a useful method for fire disaster prevention and emergency operation. The accurate simulation of moisture content of dead fuels is critical for forest fire danger forecast because it's strongly related to the ignition and fire spread. Daxinganling provides many forestry productions to China but it was also one of the highly fire-prone areas in China. Developing an accurate model for estimating hour scale dead fuel moisture content in the Daxinganling is very necessary and important. A fire remote automatic weather station was settled in Huzhong National Nature Reserve which is located in the Daxinganling larch forest to collect 10-hour time-lag fuel moisture content and meteorological factors (temperature, relative humidity and wind speed). The record frequency is half-hour per time. In this research, Fosberg model and Van Wagner model were evaluated, since they were widely used in forest fire danger forecast system and based on the equilibrium moisture content (EMC) that is easy to use. Desorption and absorption processes are the key to simulate fuel moisture content. So, this research chooses continuous observation data during August 25-28, 2010 without any precipitation to validate these two processes. The results show that both Fosberg model and Van Wagner model can simulate when desorption or absorption process begins, but all of them could not simulate well for whole process independently. Fosberg model could give better simulation at desorption stage (R²=0.96, P<0.01), and Van Wagner model does better at absorption stage (R²=0.83, P<0.01). Through analysis the potential actual variation of fuel moisture content in the desorption and absorption process, the vapour exchange coefficient is nearly to be a constant in the desorption which is coincide with Fosberg model's and the absorption's the vapour exchange coefficient seem to be variable which is similar to the Van Wagner model's. Thus, a new hour scale fuel moisture model is developed considering the difference of desorption and absorption processes: the desorption process based on Fosberg model with the Simard's EMC model and the absorption process based on Van Wagner model with the Van Wagner's EMC model. The new model use the difference between fuel moisture content and Van Wagner's absorption EMC to determine the vapor exchange process, and it could give better simulation (R²=0.77, P<0.01). But the absorption processes was underestimated in the calm weather, and the modified model simulated better (R²=0.88, P<0.01).This research could provide the technical supports for the forest fire danger forecast in Daxinganling.