Changes in C4 grass distribution and abundance are frequently observed in Quaternary, Holocene and future environmental-change scenarios. However, the factors driving these dynamics are not fully understood, and conflicting theories have been reported. In this paper, we present a very large dataset of modern altitudinal distribution profiles of C3 and C4 grasses covering the entire Neotropical Andes, which was compared with actual climate data. The results of multivariate analysis demonstrate that, in the Neotropical Andes, mean annual temperature is the main factor governing the modern altitudinal distribution of C3 and C4 grass species. The C3 and C4 grass distributions were compared with simulations based on the Lund-Potsdam-Jena dynamic global vegetation model (LPJ-DGVM), which allowed the present grass distribution to be estimated. Finally, the DGVM was employed to simulate past and future scenarios, using the IPCC's climate projections for 2100 and PMIP2 models for the Holocene Optimum (HO, 6000 years bp) and the Last Glacial Maximum (LGM, 21 000 years bp). The results were found to be significantly different from those obtained using a simple photosynthetic model. According to LPJ forced with the PMIP2 models for the LGM, during the LGM, the C4 grasses would not have reached higher altitudes than found in the present day.