Multi-scale analysis of species diversity pattern can provide important information to biodiversity conservation. Additive partition of diversity approach is used to analyzing the relative contribution of alpha diversity and beta diversity to gamma diversity of plant species in oak (Quercus liaotungensis) forest in a quadrat, slpoe positon, slope three scales hierarchical sampling system. The results shows that slope scale has the most contribution to the total species richness imply that the best scale for maintain species diversity in oak forest; While the most contribution of diversity for Shannon diversity index and Simpson diversity index is the quadrat scale, the difference is due to the dominance and rarity pattern of species in community; Beta diversity components increase with scale which may reflect the synthesis of environmental heterogeneity and dispersal effects; In each scale, the contribution of beta diversity by Shannon diversity is higher than by Simpson diversity, which is due to the rare species distribution in the community. The paper argued that additive partition of diversity is an valuable analytical framework for multi-scale analysis of species diversity pattern and is an effective approach for the forming mechanism of species pattern.