Abstract: Xixia,a key forest county in Henan province,was selected as the case study area,while spatial dataset of forest resources continuous inventory in 2003 was collected as the main information sources of data mining.Spatial hot explore,trend analysis,geographically weighted regression (GWR) and C5.0 decision tree analysis were performed on the platform of ArcGIS 9.3 and Clemintine 12.0 to mining spatial knowledge of forest biomass in Xixia county.Research shows that:(1) the spatial distribution of forest biomass is closely related with gradient of elevation,showing a tilting ladder pattern from north to south;(2) forest biomass is positively correlated with average tree height and slope,negatively with soil thickness and light brightness value;(3) the importance of the four major factors is as follows:average tree height (0.30)>light brightness value (0.24)>slope (0.23)>soil thickness (0.22).

Key words: forest resources continuous inventory, biomass, spatial knowledge mining, Xixia