The purpose of this paper was mainly to interpret the pixel missing patch of image by using the kriging interpolation technology of spatial statistics. The TM image （a resolution of 30m×30m） of the Jingouling Forest Farm of the Wangqing Forestry Bureau of Northeast China on 1 July

1997 was selected as data resource

and some patches with cloud and cloud shade from the TM image were adopted as experimental data in this paper. Based on the classification for the TM remote sensing image

the pixel information of real pixel missing patches of image （the real pixel missing patch is the patch with the unknown real information of partial regions on the image

which is caused by the shelters of cloud

cloud shade

soot and others） was restored by the kriging interpolation technology of spatial statistics theory under the image processing software—ERDAS

the geographic information system software—ArcInfo and the spatial statistics analysis software—ILWIS. The restored pixel information was reclassified

and then compared with the results of forest resource survey

thus the results （i.e. the classified precision of the restored information of each patch was more than 75%）

was obtained. Moreover

the kriging interpolation technology used in this paper grasped the fact that the objects on the global surface possessed continuity and self correlation

and therefore it had better estimation effects. Additionally

the applicable scopes and the applicable conditions of the kriging interpolation were discussed in detail from different aspects. In summary

this paper would provide a method and means for interpreting the pixel missing patch of image.