Geographical patterns of biodiversity and their underlying mechanisms have been the focus of macroecological and biogeographical studies. Many hypotheses have been proposed to explain macro-scale geographic patterns of species richness, such as the water-energy, habitat heterogeneity, and evolutionary history hypotheses. Extensive studies have shown that at landscape scales, habitat fragmentation has important effects on spatial patterns of species diversity, but it is still unclear whether these effects on geographical patterns of biodiversity are pronounced at macroecological scales. Recent research has focused on the role of local landscape attributes in driving macro-scale biodiversity patterns. However, several different studies have drawn contradictory conclusions. In the present study, we used the distribution range data of 575 mammal species and 1328 avian species in mainland China extracted from the database of China's Species Information System to generate the geographical patterns of species richness per 100 km×100 km grid cells. We further used ordinary least squares (OLS) models and simultaneous autoregressive (SAR) models to examine the relationship between species richness and climate (including 19 bioclimatic variables), habitat heterogeneity (including 20 variables), and landscape fragmentation (including 24 patch-and landscape-level metrics representing fragmentation). Results showed that the SAR models could effectively solve the spatial autocorrelation problem, as indicated by the substantially reduced Moran's I index. The landscape fragmentation factors were significantly correlated with the species richness of both birds and mammals, presenting explanatory power of 30%-50% (non-spatial OLS models) or 60%-80% (SAR models), which is slightly lower than that of the climate and habitat heterogeneity factors. This indicates that the effect of local landscape fragmentation is non-trivial regarding macro-scale patterns of species richness. We further used the variation partitioning technique to quantitatively compare the relative importance of the three groups of influencing factors, i.e., climate, habitat heterogeneity, and local landscape fragmentation. The results showed that landscape fragmentation had largely overlapped explanatory power of 20%-40% with climate and habitat heterogeneity. The explanatory power of landscape fragmentation was higher for mammal than for bird species richness. In summary, our results suggested that local landscape fragmentation plays an important role in driving macro-scale species richness patterns. Therefore, local landscape attributes that model large-scale biodiversity patterns should be considered in future research. Our research will facilitate a better understanding of the drivers of biodiversity patterns at macroecological scales, and provides important implications for biological conservation.