It is noted that forest spatial structure significantly affects the understory species diversity. In the present study, for identifying the spatial structure factors that affect understory shrub species diversity in the Quercus secondary forest, Pearson's correlation analysis, step-wise multiple linear regression models, and canonical correlation analysis (CCA) were used to explore the relationships between spatial structure and shrub species diversity. The forest spatial structure was expressed by the mingling index, aggregation index, and opening degree index. Margalef index, Auclair & Goff index, and Pielou's index were used to describe shrub biodiversity. Results showed that:(1) The mingling index had significant influence on the Pielou and Auclair & Goff index of the shrub layer (P < 0.05) according to the Pearson's correlation analysis. In contrast, the Margalef index was not significant (P > 0.05). (2) Shrub species diversity was influenced by several spatial structure indices at the same time through multiple linear regression analysis. Pielou's index was affected by the mingling index, aggregation index, and opening degree index. The Auclair & Goff index was affected by the mingling index and opening degree index. (3) Given the linear combination of spatial structure indices and understory shrub diversity indices, the canonical correlation coefficient was 0.5427. Especially, the mingling index was the most important variable affecting the shrub species diversity. Therefore, to improve the shrub species diversity of the Quercus secondary forests, it is necessary to focus on regulating the forest species structures, and following by forest spatial distributions.