Temperate steppes exhibit complex vegetation types and diverse characteristics across varied climate gradients. However, to reconstruct past changes in temperate steppe vegetation and climate using pollen data, it is essential to establish more precise modern relationships between pollen, vegetation and climate. In this paper, we report the results of detailed vegetation surveys and palynological analyses for 149 sampling sites in the Xilingol Steppe, and apply a series of statistical analysis methods, including Kriging interpolation and ordination analysis, to evaluate how modern pollen assemblages reflect vegetation and climatic gradients. Our results show that Amaranthaceae and Artemisia dominate the surface pollen assemblages of the steppe. The spatial heterogeneity of pollen assemblages is highly significant and allows us to distinguish meadow steppes, typical steppes, and desert steppes. Specifically, the percentage of Poaceae pollen serves as a key indicator for differentiating steppe types, decreasing progressively by approximately 75 % from meadow steppes to typical steppes and then to desert steppes. Cyperaceae pollen typically represents meadow steppe landscapes, whereas Allium and Tribulus pollen are indicators of desert steppes. Arboreal pollen, primarily exogenous inputs such as Pinus originating from the Greater Khingan Mountains, reflects external influences. Regional precipitation emerges as the most critical climatic factor influencing the spatial variation of surface pollen assemblages. High percentages of Amaranthaceae and Tribulus pollen indicate arid conditions, and the percentage of Cyperaceae pollen serves as an important indicator of humid climates. In summary, this study provides detailed and reliable evidence for exploring the modern processes of pollen in East Asian steppe regions and enhances the accuracy of reconstructing past environmental changes.