The poems associated with the Grand Canal Poetry Road emphasize its cultural significance and impact. Analyzing the landscapes depicted in these poems offers insights into the spatial distribution characteristics of the Grand Canal's cultural landscape. This study examines 1 115 poems during the Tang and Song Dynasties along the Grand Canal Poetry Road, employing Natural Language Processing (NLP) techniques to extract landscape elements. Geographic Information Systems (GIS) and Geodetector methods were utilized to uncover the spatiotemporal distribution patterns and driving forces behind these landscapes. The results indicate minimal differences in landscape characteristics between the Tang and Song Dynasties. The landscapes of the Tang Dynasty are described as “secluded mountains and wilderness”, while those of the Song Dynasty are denoted as “forested retreats and garden residences”. The centers of spatial distribution for poetic landscapes shifted only slightly between the two periods, forming a "two-core and multi-point" distribution pattern. Cultural factors served as the main drivers of landscape evolution in both dynasties, with economic factors taking on an increasingly important role during the Song Dynasty. Although natural factors exerted limited independent influence, their explanatory power significantly increased when interacting with cultural and economic factors. These findings affirm the feasibility of utilizing poetic texts to construct historical GIS-based spatial analyses, providing theoretical support for the study and preservation of the Grand Canal's cultural landscapes.