This paper deals with the dynamics of forest resources by using time-varing models. The varing factors to effect forest resources are used as input vectors, which derives the stochastic difference equations. The dynamics of forest resources is described as state equations and output equations.The Kalman filtering is used to forecast the dynamics of forest resources in Guangxi. The forecasting results are cheked and analyzed. The forecasting models and methods discussed in this paper are quite effective