The scarcity of precipitation and unique desert landscapes drive distinct surface processes, making the quantitative reconstruction of precipitation changes in deserts a challenge in past global change and landscape evolution studies. Establishing non-biological proxy-precipitation transfer functions not only provides new methods for quantitatively reconstructing desert hinterland environments but also offers an effective comparison and validation approach for the reliability of biological proxies. However, it remains uncertain whether the geochemistry element ratio can reliably serve as a proxy for precipitation in desert regions. In this study, 47 surface eolian sand samples were collected across the Alashan Desert in the arid region of northwestern China, and their bulk and 40 different grain size fractions (> 250 mu m and 125-250 mu m) samples were analyzed for Na2O and K2O concentrations by X-ray fluorescence spectrometry. By combining previously published data in this region, we calculated and analyzed the spatial variation of the Na/K ratio using a total of 161 bulk eolian sand samples. The results show that the spatial variation of the Na/K ratio primarily reflects mean annual precipitation (MAP), rather than being influenced by source rocks or sedimentary sorting, with the Na/K ratio decreasing as precipitation increases. Furthermore, in the arid desert regions (the Alashan Desert, Taklimakan Desert, Qaidam Desert, and Kumtagh Desert, MAP < 200 mm) of northern China, the Na/K ratio shows a strong and significant negative correlation with MAP (MAP = -156.84 x Na/K + 300.36, R-2 = 0.86, RMSE = 30 mm), indicating that Na/K ratios can provide valuable quantitative estimates of paleo-precipitation changes in the arid desert regions of northern China.