Rodent-mediated seed dispersal largely affects the regeneration and colonization of the forest vegetation. However, due to the steep topography, complex terrains and the heavy anthropogenic logging from 1970s to 1990s, the secondary succession process of the forest is greatly inhibited where temperate deciduous broadleaf forests were the zonal vegetation. Previous studies have ignored the seed dispersal limitation mechanism among different slope positions in montane forests. We established 90 sample plots in Taihang Mountains among different slope positions (i.e., ridge, midslope and valley), and investigated the characteristics of seed removal rate, seed fate and seed dispersal distance of Quercus wutaishanica forest according to three slope positions. The results showed that only one from each of the three rodent species was captured at the ridge, while 52.1% and 43.8% of the small rodents were found in valley and midslope, respectively. Compared to the ridge whose almost all released seeds were intact in site, the seed removal rates were significantly higher in midslope and valley, and the proportions of scatter hoarded in ridge and midslope were significantly different, while both has no significant difference with that in valley. The average seed dispersal distance in midslope was 4.78 m, significantly greater than that in valley, while that of the ridge was only 2.09 m. Therefore, the midslope had the best seed dispersal, but the seed dispersal of ridge was severely restricted, providing the first empirical evidence for the Mid-domain Effect model and the Resource Availability Hypothesis. These results provide a better understanding of the dispersal limitation mechanism of the oak forest and the plant-animal interactions system in mountainous areas.