Understanding how biotic and abiotic filters drive plant distributions is critical for informing restoration efforts of degraded lands. However, it remains unclear whether or not it is possible to develop efficient restoration strategies that are based on field investigations? Here, we conducted a field-based investigation along a land degradation gradient on the Tibetan Plateau. Along this gradient, we compared species composition, Pianka's niche overlap, and total aboveground biomass among grasses, sedges, legumes, and forbs. We also used Faith's (1992) phylogenetic diversity index (PD) and the net relatedness index (NRI) to measure community phylogenetic patterns. To understand the effects of abiotic filtering, we measured soil pH, moisture, electrical conductivity, total soil organic carbon, microbial biomass carbon, and microbial biomass nitrogen. Our results show that grasses replaced sedges and mesoxerophytes replaced mesohydrophytes with increasing land degradation. We also found that niche overlap declined with increasing degradation intensity. Furthermore, grass biomass increased, while sedges and legumes decreased, and forb biomass showed a hump-shaped relationship with degradation intensity. In addition, the communities transitioned from a clustered to a more random phylogenetic pattern. Soil conditions changed from acidic to alkaline and from fertile to harsh (p < .05). High niche overlap indicates a high biotic filtering at lightly degraded land where abiotic filtering is less stressful. However, when degradation is intense, abiotic filtering is apparent, as indicated by low levels of niche overlap. Overall, we suggest that distantly related, arid tolerant and hardy plants, and soil reclamation should be considered in degraded alpine lands restoration.