Abstract: Based on the 2007 Landsat TM images and the dominant environmental factors of shelter forest, the forest sites in the Three Gorges Reservoir Area were classified, and by using multiple objective grey situation decision model, three indices including water conservation amount, biomass, and stand productivity were selected to make the spatial optimum allocation of the present four kinds of shelter forest (coniferous forest, broadleaf forest, mixed coniferousbroadleaf forest, and shrub) in the Area. The forest sites in the Area in 2007 could be classified into 40 types, and after the optimization of spatial allocation, the proportion of coniferous forest, broadleaf forest, mixed coniferous-broadleaf forest, and shrub would be 32.55%, 29.43%, 34.95%, and 3.07%, respectively. Comparing with that before optimization, the proportion of coniferous forest and shrub after optimization was reduced by 8.79% and 28.55%, while that of broadleaf forest and mixed coniferous-broadleaf forest was increased by 10.23% and 27.11%, respectively. After the optimization of spatial allocation, the amount of water conservation, biomass, and stand productivity of the shelter forests in the area would be increased by 14.09×10⁸ m³, 0.35×10⁸ t, and 1.08×10⁶ t, respectively.

Key words: Three Gorges Reservoir Area, shelter forest, spatial optimum allocation, integrated benefit, multiple objective grey situation decision, Reaumuria soongorica, early selection, superior family, principal component analysis, membership function method.