Abstract: Pollen and phytolith are valuable proxies for reconstructing ancient agricultural landscapes and environmental changes. Understanding their distribution patterns in the topsoil and their relationship with vegetation composition is vital for accurately interpreting fossil pollen and phytolith data. We analyzed pollen and phytolith data from 22 topsoil samples representative of three land-use types in Langqi Island, Fujian Province. We presented the spatial distributions of different pollen taxa and their quantitative relationship with fractional vegetation cover (FVC) based on geographic information system and remote sensing techniques. Pollen analysis revealed that Pinus pollen, as an exotic species, were pervasive in the samples, which diluted the representation of local components, and generated obvious bias between pollen assemblages and vegetation composition of local agricultural lands. Furthermore, agricultural harvesting decreased the pollen representation of cultivated plants. The FVC extracted from remote sensing images is linearly correlated with the abundance of arboreal pollen in the uncultivated land (R²=0.7764), implying that dense herbs on the uncultivated land decreased the share of exotic arboreal pollen, and confirming excellent pollen representation of this vegetation type. Phytolith analysis indicated that numerous phytoliths occurred in the samples, such as Bambusoideae, Oryza sativa, and weeds; they provided more detailed information on non-arboreal plants than pollen proxy.