This paper used the typical sampling in order to study the impact of four types of thinning intensities (19%, 33%, 55%, and 62%) on the species density, diversity and coverage of undergrowth vegetation, and provide a theoretical basis for the management in natural forest of Altai Mountains. In this study, taking the natural forest of Altai Mountains as object with 12 samples were set up to measure all trees of different sample plots after 30 years thinning, in the same time, undergrowth vegetation structure (coverage and density), plant composition, and diversity(richness index, Shannon-Wiener index, Simpson index, and Pielou index) in the disturbed communities were investigated. One-way analysis of variance (ANOVA) and Fisher's list significant difference (LSD) multiple comparisons were used for significance test and Pearson correlation analysis among under-story vegetation was conducted. The investigation of the plot showed that:(1) appropriate thinning improved the density of under-story plants. On the one hand, herb species richness increasingly improved with an increasing thinning intensity and it was significantly higher than that of other treatment at 62% thinning intensity. The species richness of shrub increased first and then decreased with increasing thinning intensity, which reached the highest with 55% thinning intensity layer, but the different thinning intensities had no significant effect on the shrub layer species richness. On the other hand, Simpson index, Shannon-Wiener index and Pielou evenness index, the thinning intensity value of each undergrowth vegetation layers reached the highest when with 55% thinning layers, while the different thinning intensities had no significant effects among the shrub layers in terms of Margarlef richness index, Species richness index, Shannon-Wiener index, Simpson index and Pielou index and herb layer in terms of Species richness index, Shannon-Wiener index, Simpson index and Pielou index with different thinning intensities treatment (P>0.05). The dominant species in herb layer was Poanemoralis L., which is similar with those among 4 different thinning intensities. (2) Appropriate increase of thinning intensity significantly promoted the density of shrub and herb layer under forest, and the result shows that 55% thinning intensity layer was the best (P<0.05). The shrub and herb layer coverage was increased first and then decreased with increasing thinning intensity, the increase of thinning intensity had obviously change on the coverage of shrub layer, while it had no significant effect on herb layer coverage (P>0.05). In conclusion, from the perspective of restoring forest land diversity and forest ecosystem stability, 55% thinning intensity is more reasonable than others.