The segregation relationship between different species pairs is helpful for revealing the phenomena of species interactions, community structure and dynamics. It can be used to explore the impacts of environmental differences on species distribution. Practically, such study is the scientific fundamental of species conservation and restoration and reconstruction of degraded ecosystems. The Karst forest in Maolan National Natural Reserve of Guizhou Province, southwestern China is a rare, original forest remnant in the mid-subtropics of the world. However, little is known about the interspecific relationship of the species. In order to investigate the interspecific segregation characteristics of the species in karst forest, so as to discuss the mechanism of species coexistence on Karst habitat,We set up two plots of 1hm2 (named P1 and P2, respectively) in two old-growth, evergreen-deciduous broadleaved mixed Karst forests in Maolan National Nature Reserve, Guizhou Province of SW China. All woody species with the diameter at breast height (DBH) ≥1cm were identified and surveyed. The distribution map of 43 tree species with abundance ≥20 individuals from two Karst forest communities were drawn using the Geographical Information System software (ArcGIS). The nearest neighbors of each individual and the distances between every individual-neighbor pair were obtained by using ArcGIS’s extension module. The interspecific segregation in the multi-species community was then analyzed using a sub-table method of a N×N nearest-neighbor contingency table. The Pielou′s coefficient of segregation (S) was adopted to measure the segregated degree between two species. The results show that segregation relations between most of the species are random in two karst forest communities (60.91% and 63.34%, respectively). Some species pairs are positively segregated (38.76% and 3566%, respectively) and a few pairs are negatively segregated (0.33% and 1.0%, respectively). Our results support the Pielou′s opinion that negative segregation is rare in old-growth plant communities. This showed that interspecific competitive exclusion is weakened between the species, and the two forests were at a stable climax stage. According to their adaptability to the environment and relationships between plants and environment, the 43 species from two forest plots were divided into three ecological species groups (ESG). In Plot one, the first ESG includes 17 big tree species such as Cyclobalanopsis glauca, Clausena dunniana, and Engelhardtia roxburghiana, which are light-demanding and drought-tolerant in the first canopy tree layer. The second ESG is 18 tree species in the sub-canopy tree layer including Acer wangchii, Viburnum brachybotryum, and Swida parviflora etc that have less drought-tolerance and some extent shade-tolerance. The ESG 3 has 8 woody species with broad distribution range in mostly the shrub layer. In Plot two the ESG 1 has 13 dominate tree species including Castanopsis carlesii var. spinulosa, Rapanea neriifolia, and Osmanthus fragrans etc in habitat of relatively high soil moisture. 19 tree species such as Distylium myricoides, Lindera communis, and Clausena dunniana occur in the ESG 2. These species grow in relatively dry habitat. The ESG 3 has 12 tree species of sub-canopy layer, occupying broader ecological niche in Plot two. The ecological characteristics and habitat preference of the species in each group showed significant differences. The results also indicated that interspecific segregation is closely related to habitat heterogeneity of karst terrain. Species pairs in different habitats mostly tend to be more positively segregated, but species with similar habitat preference or widely distributed are more likely to be random neighbor or negatively segregated. Therefore, interspecific relationship based on differentiation of habitat may be one of the important mechanisms of species coexistence in species-rich karst forest.