The present study compares the community dynamics of spiders in the rubber plantation of Hainan Island to those of spiders in the island's natural forest. Hainan, which is the fifth largest rubber producer in the world, is the second largest island in China. It features a tropical climate, and is located in the South China Sea. Rubber plantations are commonly considered "Green Deserts", as the expansion of rubber plantations has caused a series of ecological problems, including habitat fragmentation, biodiversity loss, habitat loss, soil erosion, and climate change. Of the arthropods, spiders comprise a group known to measure changes in habitat structure, habitat type, wind, and temperature exposure, and play key ecological roles as predators. Spider assemblages, which are easily sampled, are useful indicators that can be used to compare the biodiversity of various environments and assess the effects of disturbances on diversity. Because spiders are generally found in forests in high abundance, they form a good model taxa for biodiversity studies. In addition, their distributions and abundances are linked to the structural attributes of their habitat. They play key functional roles in ecosystems, and their abundance, richness, and community structure in tree canopies are associated with the complexity of the ecosystem's vegetative structure. To compare the composition and diversity of spiders in a natural forest and a rubber plantation, spider communities in 6 plots (three repetitions for each forest type) were investigated during four field trips in August 2010, using pitfall traps, direct searching, and sweep-netting. The similarity analysis of samples from the natural forest and rubber plantation was conducted via the principal component analysis (PCA) of spider families and numbers of individuals. A total of 3609 individual spiders were collected from the two forests, 969 of which were identifiable adults representing 23 families and 162 species. From samplings of the natural forest, 20 families and 100 species were identified, in which the most abundant families were Salticidae, Theridiidae, and Araneidae. From the rubber plantation samples, 17 families and 87 species were identified, in which the most abundant families were Tetragnathidae, Lycosidae, and Oxyopidae. The abundance of spiders in the rubber plantation is significantly higher than that in the natural forest, but the Shannon -Wiener (H'), Simpson index (D), Evenness (J), and Richness (D_mg) are significantly higher in the natural forest than in the rubber plantation. In the rubber plantation, the abundance of orb weavers and cursorial hunter-spiders is significantly higher than that in the natural forest, while no significant difference can be detected between sheet-line weaver spiders and ambush predator spiders in the two forests. The results of the PCA indicate that the natural forest and rubber plantation are highly differentiated, with the sampling plots in the natural forest having a much higher degree of similarity than those in the rubber plantation. These results suggest that (1) spider community structure varies and its diversity decreases when rubber plantations replace natural forest, (2) increasing the complexity of habitat structure and reducing anthropogenic disturbance is a meaningful way to conserve and restore biodiversity.