Northern Shaanxi is important in understanding the ancient use and northward spread of foxtail millet (Setaria italica) and broomcorn millet (Panicum miliaceum). Nonetheless, due to the lack of millet remains, AMS radiocarbon data, and environmental background, the emergence, crop structure, and environmental factors of millet use in northern Shaanxi remain ambiguous. To address this knowledge gap, a systematic survey was conducted along the Beiluo River. Forty-two relic units at 19 Neolithic sites were selected for analysis through phytolith, AMS radiocarbon dating, and spatio-temporal approaches. Phytolith and AMS radiocarbon dating analyses traced the utilization of millets in the Beiluo River to 6280 cal. BP. In addition, broomcorn millet was more prevalent than foxtail millet during the Neolithic period, although the prevalence of the latter increased during the late Longshan period. Spatio-temporal analysis demonstrated that millets initially appeared in the Beiluo River during the Yangshao period, gradually moving away during the Longshan period, which was probably first related to the nearest rivers and then the spread of cattle and sheep. However, the millet cultivation altitude remained at 1400 m throughout the Yangshao and Longshan periods. Collectively, these findings provide evidence for the use and northward spread of millets in northwest China.

Over the past decade, the rapid development of geospatial tools has significantly expanded the scope of archaeobotanical research, enabling unprecedented insights into ancient plant domestication, agricultural practices, and human-environment interactions. Within the Chinese context, where rich archaeobotanical records intersect with complex socio-ecological histories, GIS-driven approaches have revealed nuanced patterns of crop dispersal, settlement dynamics, and landscape modification. However, despite these advances, current applications remain largely exploratory, constrained by fragmented datasets and underutilized spatial-statistical methods. This paper argues that a more robust integration of large-scale archaeobotanical datasets with advanced ArcGIS functionalities—such as kernel density estimation, least-cost path analysis, and predictive modelling—is essential to address persistent gaps in the field. By synthesizing case studies from key Chinese Neolithic and Bronze Age sites, we demonstrate how spatial analytics can elucidate (1) spatiotemporal trends in plant use, (2) anthropogenic impacts on vegetation, and (3) the feedback loops between subsistence strategies and landscape evolution. Furthermore, we highlight the challenges of data standardization, scale dependency, and interdisciplinary collaboration in archaeobotanical ArcGIS. Ultimately, this study underscores the imperative for methodological harmonization and computational innovation to unravel the intricate relationships between ancient societies, agroecological systems, and long-term environmental change.